From 05209a43923001ca04acee9ded9d4fc7a593f9ee Mon Sep 17 00:00:00 2001
From: Jiuqiang Tang <jqtang@google.com>
Date: Tue, 4 Oct 2022 04:36:56 -0700
Subject: [PATCH 1/5] Refactor mediapipe_aar.bzl to expose
 `mediapipe_java_proto_srcs`, `mediapipe_logging_java_proto_srcs`, and
 `mediapipe_java_proto_src_extractor`.

PiperOrigin-RevId: 478750184
---
 .../com/google/mediapipe/mediapipe_aar.bzl    | 194 +++++++++---------
 1 file changed, 92 insertions(+), 102 deletions(-)

diff --git a/mediapipe/java/com/google/mediapipe/mediapipe_aar.bzl b/mediapipe/java/com/google/mediapipe/mediapipe_aar.bzl
index ed1686954..7f2cb146c 100644
--- a/mediapipe/java/com/google/mediapipe/mediapipe_aar.bzl
+++ b/mediapipe/java/com/google/mediapipe/mediapipe_aar.bzl
@@ -89,10 +89,6 @@ def mediapipe_aar(
         calculators = calculators,
     )
 
-    _mediapipe_proto(
-        name = name + "_proto",
-    )
-
     native.genrule(
         name = name + "_aar_manifest_generator",
         outs = ["AndroidManifest.xml"],
@@ -115,19 +111,10 @@ EOF
                    "//mediapipe/java/com/google/mediapipe/components:java_src",
                    "//mediapipe/java/com/google/mediapipe/framework:java_src",
                    "//mediapipe/java/com/google/mediapipe/glutil:java_src",
-                   "com/google/mediapipe/formats/annotation/proto/RasterizationProto.java",
-                   "com/google/mediapipe/formats/proto/ClassificationProto.java",
-                   "com/google/mediapipe/formats/proto/DetectionProto.java",
-                   "com/google/mediapipe/formats/proto/LandmarkProto.java",
-                   "com/google/mediapipe/formats/proto/LocationDataProto.java",
-                   "com/google/mediapipe/proto/CalculatorProto.java",
-               ] +
+               ] + mediapipe_java_proto_srcs() +
                select({
                    "//conditions:default": [],
-                   "enable_stats_logging": [
-                       "com/google/mediapipe/proto/MediaPipeLoggingProto.java",
-                       "com/google/mediapipe/proto/MediaPipeLoggingEnumsProto.java",
-                   ],
+                   "enable_stats_logging": mediapipe_logging_java_proto_srcs(),
                }),
         manifest = "AndroidManifest.xml",
         proguard_specs = ["//mediapipe/java/com/google/mediapipe/framework:proguard.pgcfg"],
@@ -179,93 +166,6 @@ EOF
 
     _aar_with_jni(name, name + "_android_lib")
 
-def _mediapipe_proto(name):
-    """Generates MediaPipe java proto libraries.
-
-    Args:
-      name: the name of the target.
-    """
-    _proto_java_src_generator(
-        name = "mediapipe_log_extension_proto",
-        proto_src = "mediapipe/util/analytics/mediapipe_log_extension.proto",
-        java_lite_out = "com/google/mediapipe/proto/MediaPipeLoggingProto.java",
-        srcs = ["//mediapipe/util/analytics:protos_src"],
-    )
-
-    _proto_java_src_generator(
-        name = "mediapipe_logging_enums_proto",
-        proto_src = "mediapipe/util/analytics/mediapipe_logging_enums.proto",
-        java_lite_out = "com/google/mediapipe/proto/MediaPipeLoggingEnumsProto.java",
-        srcs = ["//mediapipe/util/analytics:protos_src"],
-    )
-
-    _proto_java_src_generator(
-        name = "calculator_proto",
-        proto_src = "mediapipe/framework/calculator.proto",
-        java_lite_out = "com/google/mediapipe/proto/CalculatorProto.java",
-        srcs = ["//mediapipe/framework:protos_src"],
-    )
-
-    _proto_java_src_generator(
-        name = "landmark_proto",
-        proto_src = "mediapipe/framework/formats/landmark.proto",
-        java_lite_out = "com/google/mediapipe/formats/proto/LandmarkProto.java",
-        srcs = ["//mediapipe/framework/formats:protos_src"],
-    )
-
-    _proto_java_src_generator(
-        name = "rasterization_proto",
-        proto_src = "mediapipe/framework/formats/annotation/rasterization.proto",
-        java_lite_out = "com/google/mediapipe/formats/annotation/proto/RasterizationProto.java",
-        srcs = ["//mediapipe/framework/formats/annotation:protos_src"],
-    )
-
-    _proto_java_src_generator(
-        name = "location_data_proto",
-        proto_src = "mediapipe/framework/formats/location_data.proto",
-        java_lite_out = "com/google/mediapipe/formats/proto/LocationDataProto.java",
-        srcs = [
-            "//mediapipe/framework/formats:protos_src",
-            "//mediapipe/framework/formats/annotation:protos_src",
-        ],
-    )
-
-    _proto_java_src_generator(
-        name = "detection_proto",
-        proto_src = "mediapipe/framework/formats/detection.proto",
-        java_lite_out = "com/google/mediapipe/formats/proto/DetectionProto.java",
-        srcs = [
-            "//mediapipe/framework/formats:protos_src",
-            "//mediapipe/framework/formats/annotation:protos_src",
-        ],
-    )
-
-    _proto_java_src_generator(
-        name = "classification_proto",
-        proto_src = "mediapipe/framework/formats/classification.proto",
-        java_lite_out = "com/google/mediapipe/formats/proto/ClassificationProto.java",
-        srcs = [
-            "//mediapipe/framework/formats:protos_src",
-        ],
-    )
-
-def _proto_java_src_generator(name, proto_src, java_lite_out, srcs = []):
-    native.genrule(
-        name = name + "_proto_java_src_generator",
-        srcs = srcs + [
-            "@com_google_protobuf//:lite_well_known_protos",
-        ],
-        outs = [java_lite_out],
-        cmd = "$(location @com_google_protobuf//:protoc) " +
-              "--proto_path=. --proto_path=$(GENDIR) " +
-              "--proto_path=$$(pwd)/external/com_google_protobuf/src " +
-              "--java_out=lite:$(GENDIR) " + proto_src + " && " +
-              "mv $(GENDIR)/" + java_lite_out + " $$(dirname $(location " + java_lite_out + "))",
-        tools = [
-            "@com_google_protobuf//:protoc",
-        ],
-    )
-
 def _mediapipe_jni(name, gen_libmediapipe, calculators = []):
     """Generates MediaPipe jni library.
 
@@ -345,3 +245,93 @@ cp -r lib jni
 zip -r $$origdir/$(location :{}.aar) jni/*/*.so
 """.format(android_library, name, name, name, name),
     )
+
+def mediapipe_java_proto_src_extractor(target, src_out, name = ""):
+    """Extracts the generated MediaPipe java proto source code from the target.
+
+    Args:
+      target: The java proto lite target to be built and extracted.
+      src_out: The output java proto src code path.
+      name: The optional bazel target name.
+
+    Returns:
+      The output java proto src code path.
+    """
+
+    if not name:
+        name = target.split(":")[-1] + "_proto_java_src_extractor"
+    src_jar = target.replace("_java_proto_lite", "_proto-lite-src.jar").replace(":", "/").replace("//", "")
+    native.genrule(
+        name = name + "_proto_java_src_extractor",
+        srcs = [target],
+        outs = [src_out],
+        cmd = "unzip $(GENDIR)/" + src_jar + " -d $(GENDIR) && mv $(GENDIR)/" +
+              src_out + " $$(dirname $(location " + src_out + "))",
+    )
+    return src_out
+
+def mediapipe_java_proto_srcs(name = ""):
+    """Extracts the generated MediaPipe framework java proto source code.
+
+    Args:
+      name: The optional bazel target name.
+
+    Returns:
+      The list of the extrated MediaPipe java proto source code.
+    """
+
+    proto_src_list = []
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/framework:calculator_java_proto_lite",
+        src_out = "com/google/mediapipe/proto/CalculatorProto.java",
+    ))
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/framework/formats:landmark_java_proto_lite",
+        src_out = "com/google/mediapipe/formats/proto/LandmarkProto.java",
+    ))
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/framework/formats/annotation:rasterization_java_proto_lite",
+        src_out = "com/google/mediapipe/formats/annotation/proto/RasterizationProto.java",
+    ))
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/framework/formats:location_data_java_proto_lite",
+        src_out = "com/google/mediapipe/formats/proto/LocationDataProto.java",
+    ))
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/framework/formats:detection_java_proto_lite",
+        src_out = "com/google/mediapipe/formats/proto/DetectionProto.java",
+    ))
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/framework/formats:classification_java_proto_lite",
+        src_out = "com/google/mediapipe/formats/proto/ClassificationProto.java",
+    ))
+    return proto_src_list
+
+def mediapipe_logging_java_proto_srcs(name = ""):
+    """Extracts the generated logging-related MediaPipe java proto source code.
+
+    Args:
+      name: The optional bazel target name.
+
+    Returns:
+      The list of the extrated MediaPipe logging-related java proto source code.
+    """
+
+    proto_src_list = []
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/util/analytics:mediapipe_log_extension_java_proto_lite",
+        src_out = "com/google/mediapipe/proto/MediaPipeLoggingProto.java",
+    ))
+
+    proto_src_list.append(mediapipe_java_proto_src_extractor(
+        target = "//mediapipe/util/analytics:mediapipe_logging_enums_java_proto_lite",
+        src_out = "com/google/mediapipe/proto/MediaPipeLoggingEnumsProto.java",
+    ))
+    return proto_src_list

From 8d5cf9bbedb2ee97651f57feeb273f0da48e4f78 Mon Sep 17 00:00:00 2001
From: Sebastian Schmidt <mrschmidt@google.com>
Date: Tue, 4 Oct 2022 09:40:03 -0700
Subject: [PATCH 2/5] Open source MediaPipe object detector task reference app
 prototype.

PiperOrigin-RevId: 478811683
---
 .../instantmotiontracking/GIFEditText.java    |   2 +-
 mediapipe/tasks/examples/android/BUILD        |  21 ++
 .../src/main/AndroidManifest.xml              |  37 +++
 .../android/objectdetector/src/main/BUILD     |  48 ++++
 .../examples/objectdetector/MainActivity.java | 236 ++++++++++++++++++
 .../ObjectDetectionResultImageView.java       |  77 ++++++
 .../drawable-v24/ic_launcher_foreground.xml   |  34 +++
 .../res/drawable/ic_launcher_background.xml   |  74 ++++++
 .../android/res/layout/activity_main.xml      |  40 +++
 .../res/mipmap-anydpi-v26/ic_launcher.xml     |   5 +
 .../mipmap-anydpi-v26/ic_launcher_round.xml   |   5 +
 .../android/res/mipmap-hdpi/ic_launcher.png   | Bin 0 -> 1354 bytes
 .../mipmap-hdpi/ic_launcher_foreground.png    | Bin 0 -> 2257 bytes
 .../res/mipmap-hdpi/ic_launcher_round.png     | Bin 0 -> 3246 bytes
 .../android/res/mipmap-mdpi/ic_launcher.png   | Bin 0 -> 959 bytes
 .../mipmap-mdpi/ic_launcher_foreground.png    | Bin 0 -> 900 bytes
 .../res/mipmap-mdpi/ic_launcher_round.png     | Bin 0 -> 1955 bytes
 .../android/res/mipmap-xhdpi/ic_launcher.png  | Bin 0 -> 1971 bytes
 .../mipmap-xhdpi/ic_launcher_foreground.png   | Bin 0 -> 1845 bytes
 .../res/mipmap-xhdpi/ic_launcher_round.png    | Bin 0 -> 4658 bytes
 .../android/res/mipmap-xxhdpi/ic_launcher.png | Bin 0 -> 3562 bytes
 .../mipmap-xxhdpi/ic_launcher_foreground.png  | Bin 0 -> 5655 bytes
 .../res/mipmap-xxhdpi/ic_launcher_round.png   | Bin 0 -> 7745 bytes
 .../res/mipmap-xxxhdpi/ic_launcher.png        | Bin 0 -> 5004 bytes
 .../mipmap-xxxhdpi/ic_launcher_foreground.png | Bin 0 -> 8278 bytes
 .../res/mipmap-xxxhdpi/ic_launcher_round.png  | Bin 0 -> 11062 bytes
 .../examples/android/res/values/colors.xml    |   6 +
 .../examples/android/res/values/strings.xml   |   6 +
 .../examples/android/res/values/styles.xml    |  11 +
 29 files changed, 601 insertions(+), 1 deletion(-)
 create mode 100644 mediapipe/tasks/examples/android/BUILD
 create mode 100644 mediapipe/tasks/examples/android/objectdetector/src/main/AndroidManifest.xml
 create mode 100644 mediapipe/tasks/examples/android/objectdetector/src/main/BUILD
 create mode 100644 mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/MainActivity.java
 create mode 100644 mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/ObjectDetectionResultImageView.java
 create mode 100644 mediapipe/tasks/examples/android/res/drawable-v24/ic_launcher_foreground.xml
 create mode 100644 mediapipe/tasks/examples/android/res/drawable/ic_launcher_background.xml
 create mode 100644 mediapipe/tasks/examples/android/res/layout/activity_main.xml
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher.xml
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher_round.xml
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher_foreground.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher_round.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher_foreground.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher_round.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher_foreground.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher_round.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher_foreground.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher_round.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher_foreground.png
 create mode 100644 mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher_round.png
 create mode 100644 mediapipe/tasks/examples/android/res/values/colors.xml
 create mode 100644 mediapipe/tasks/examples/android/res/values/strings.xml
 create mode 100644 mediapipe/tasks/examples/android/res/values/styles.xml

diff --git a/mediapipe/examples/android/src/java/com/google/mediapipe/apps/instantmotiontracking/GIFEditText.java b/mediapipe/examples/android/src/java/com/google/mediapipe/apps/instantmotiontracking/GIFEditText.java
index 1b733ed82..10e6422ba 100644
--- a/mediapipe/examples/android/src/java/com/google/mediapipe/apps/instantmotiontracking/GIFEditText.java
+++ b/mediapipe/examples/android/src/java/com/google/mediapipe/apps/instantmotiontracking/GIFEditText.java
@@ -18,7 +18,7 @@ import android.content.ClipDescription;
 import android.content.Context;
 import android.net.Uri;
 import android.os.Bundle;
-import androidx.appcompat.widget.AppCompatEditText;
+import android.support.v7.widget.AppCompatEditText;
 import android.util.AttributeSet;
 import android.util.Log;
 import android.view.inputmethod.EditorInfo;
diff --git a/mediapipe/tasks/examples/android/BUILD b/mediapipe/tasks/examples/android/BUILD
new file mode 100644
index 000000000..c07af2d2c
--- /dev/null
+++ b/mediapipe/tasks/examples/android/BUILD
@@ -0,0 +1,21 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+licenses(["notice"])
+
+filegroup(
+    name = "resource_files",
+    srcs = glob(["res/**"]),
+    visibility = ["//mediapipe/tasks/examples/android:__subpackages__"],
+)
diff --git a/mediapipe/tasks/examples/android/objectdetector/src/main/AndroidManifest.xml b/mediapipe/tasks/examples/android/objectdetector/src/main/AndroidManifest.xml
new file mode 100644
index 000000000..5c53dc269
--- /dev/null
+++ b/mediapipe/tasks/examples/android/objectdetector/src/main/AndroidManifest.xml
@@ -0,0 +1,37 @@
+<?xml version="1.0" encoding="utf-8"?>
+<manifest xmlns:android="http://schemas.android.com/apk/res/android"
+    package="com.google.mediapipe.tasks.examples.objectdetector">
+
+  <uses-sdk
+      android:minSdkVersion="28"
+      android:targetSdkVersion="30" />
+
+  <!-- For loading images from gallery -->
+  <uses-permission android:name="android.permission.READ_EXTERNAL_STORAGE" />
+  <!-- For using the camera -->
+  <uses-permission android:name="android.permission.CAMERA" />
+  <uses-feature android:name="android.hardware.camera" />
+  <!-- For logging solution events -->
+  <uses-permission android:name="android.permission.INTERNET" />
+  <uses-permission android:name="android.permission.ACCESS_NETWORK_STATE" />
+
+  <application
+      android:allowBackup="true"
+      android:icon="@mipmap/ic_launcher"
+      android:label="MediaPipe Tasks Object Detector"
+      android:roundIcon="@mipmap/ic_launcher_round"
+      android:supportsRtl="true"
+      android:theme="@style/AppTheme"
+      android:exported="false">
+    <activity android:name=".MainActivity"
+              android:screenOrientation="portrait"
+              android:exported="true">
+      <intent-filter>
+        <action android:name="android.intent.action.MAIN" />
+
+        <category android:name="android.intent.category.LAUNCHER" />
+      </intent-filter>
+    </activity>
+  </application>
+
+</manifest>
diff --git a/mediapipe/tasks/examples/android/objectdetector/src/main/BUILD b/mediapipe/tasks/examples/android/objectdetector/src/main/BUILD
new file mode 100644
index 000000000..65b98d647
--- /dev/null
+++ b/mediapipe/tasks/examples/android/objectdetector/src/main/BUILD
@@ -0,0 +1,48 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#      http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+licenses(["notice"])
+
+package(default_visibility = ["//visibility:private"])
+
+android_binary(
+    name = "objectdetector",
+    srcs = glob(["**/*.java"]),
+    assets = [
+        "//mediapipe/tasks/testdata/vision:test_models",
+    ],
+    assets_dir = "",
+    custom_package = "com.google.mediapipe.tasks.examples.objectdetector",
+    manifest = "AndroidManifest.xml",
+    manifest_values = {
+        "applicationId": "com.google.mediapipe.tasks.examples.objectdetector",
+    },
+    multidex = "native",
+    resource_files = ["//mediapipe/tasks/examples/android:resource_files"],
+    deps = [
+        "//mediapipe/java/com/google/mediapipe/framework/image",
+        "//mediapipe/tasks/java/com/google/mediapipe/tasks/components/containers:detection",
+        "//mediapipe/tasks/java/com/google/mediapipe/tasks/core",
+        "//mediapipe/tasks/java/com/google/mediapipe/tasks/vision/core",
+        "//mediapipe/tasks/java/com/google/mediapipe/tasks/vision/objectdetector",
+        "//third_party:androidx_appcompat",
+        "//third_party:androidx_constraint_layout",
+        "//third_party:opencv",
+        "@maven//:androidx_activity_activity",
+        "@maven//:androidx_concurrent_concurrent_futures",
+        "@maven//:androidx_exifinterface_exifinterface",
+        "@maven//:androidx_fragment_fragment",
+        "@maven//:com_google_guava_guava",
+    ],
+)
diff --git a/mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/MainActivity.java b/mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/MainActivity.java
new file mode 100644
index 000000000..7f7ec1389
--- /dev/null
+++ b/mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/MainActivity.java
@@ -0,0 +1,236 @@
+// Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package com.google.mediapipe.tasks.examples.objectdetector;
+
+import android.content.Intent;
+import android.graphics.Bitmap;
+import android.graphics.Matrix;
+import android.media.MediaMetadataRetriever;
+import android.os.Bundle;
+import android.provider.MediaStore;
+import androidx.appcompat.app.AppCompatActivity;
+import android.util.Log;
+import android.view.View;
+import android.widget.Button;
+import android.widget.FrameLayout;
+import androidx.activity.result.ActivityResultLauncher;
+import androidx.activity.result.contract.ActivityResultContracts;
+import androidx.exifinterface.media.ExifInterface;
+// ContentResolver dependency
+import com.google.mediapipe.framework.image.BitmapImageBuilder;
+import com.google.mediapipe.framework.image.Image;
+import com.google.mediapipe.tasks.core.BaseOptions;
+import com.google.mediapipe.tasks.vision.core.RunningMode;
+import com.google.mediapipe.tasks.vision.objectdetector.ObjectDetectionResult;
+import com.google.mediapipe.tasks.vision.objectdetector.ObjectDetector;
+import com.google.mediapipe.tasks.vision.objectdetector.ObjectDetector.ObjectDetectorOptions;
+import java.io.IOException;
+import java.io.InputStream;
+
+/** Main activity of MediaPipe Task Object Detector reference app. */
+public class MainActivity extends AppCompatActivity {
+  private static final String TAG = "MainActivity";
+  private static final String MODEL_FILE = "coco_ssd_mobilenet_v1_1.0_quant_2018_06_29.tflite";
+
+  private ObjectDetector objectDetector;
+
+  private enum InputSource {
+    UNKNOWN,
+    IMAGE,
+    VIDEO,
+    CAMERA,
+  }
+
+  private InputSource inputSource = InputSource.UNKNOWN;
+
+  // Image mode demo component.
+  private ActivityResultLauncher<Intent> imageGetter;
+  // Video mode demo component.
+  private ActivityResultLauncher<Intent> videoGetter;
+  private ObjectDetectionResultImageView imageView;
+
+  @Override
+  protected void onCreate(Bundle savedInstanceState) {
+    super.onCreate(savedInstanceState);
+    setContentView(R.layout.activity_main);
+    setupImageModeDemo();
+    setupVideoModeDemo();
+    // TODO: Adds live camera demo.
+  }
+
+  /** Sets up the image mode demo. */
+  private void setupImageModeDemo() {
+    imageView = new ObjectDetectionResultImageView(this);
+    // The Intent to access gallery and read images as bitmap.
+    imageGetter =
+        registerForActivityResult(
+            new ActivityResultContracts.StartActivityForResult(),
+            result -> {
+              Intent resultIntent = result.getData();
+              if (resultIntent != null) {
+                if (result.getResultCode() == RESULT_OK) {
+                  Bitmap bitmap = null;
+                  try {
+                    bitmap =
+                        downscaleBitmap(
+                            MediaStore.Images.Media.getBitmap(
+                                this.getContentResolver(), resultIntent.getData()));
+                  } catch (IOException e) {
+                    Log.e(TAG, "Bitmap reading error:" + e);
+                  }
+                  try {
+                    InputStream imageData =
+                        this.getContentResolver().openInputStream(resultIntent.getData());
+                    bitmap = rotateBitmap(bitmap, imageData);
+                  } catch (IOException e) {
+                    Log.e(TAG, "Bitmap rotation error:" + e);
+                  }
+                  if (bitmap != null) {
+                    Image image = new BitmapImageBuilder(bitmap).build();
+                    ObjectDetectionResult detectionResult = objectDetector.detect(image);
+                    imageView.setData(image, detectionResult);
+                    runOnUiThread(() -> imageView.update());
+                  }
+                }
+              }
+            });
+    Button loadImageButton = findViewById(R.id.button_load_picture);
+    loadImageButton.setOnClickListener(
+        v -> {
+          if (inputSource != InputSource.IMAGE) {
+            createObjectDetector(RunningMode.IMAGE);
+            this.inputSource = InputSource.IMAGE;
+            updateLayout();
+          }
+          // Reads images from gallery.
+          Intent pickImageIntent = new Intent(Intent.ACTION_PICK);
+          pickImageIntent.setDataAndType(MediaStore.Images.Media.INTERNAL_CONTENT_URI, "image/*");
+          imageGetter.launch(pickImageIntent);
+        });
+  }
+
+  /** Sets up the video mode demo. */
+  private void setupVideoModeDemo() {
+    imageView = new ObjectDetectionResultImageView(this);
+    // The Intent to access gallery and read a video file.
+    videoGetter =
+        registerForActivityResult(
+            new ActivityResultContracts.StartActivityForResult(),
+            result -> {
+              Intent resultIntent = result.getData();
+              if (resultIntent != null) {
+                if (result.getResultCode() == RESULT_OK) {
+                  MediaMetadataRetriever metaRetriever = new MediaMetadataRetriever();
+                  metaRetriever.setDataSource(this, resultIntent.getData());
+                  long duration =
+                      Long.parseLong(
+                          metaRetriever.extractMetadata(
+                              MediaMetadataRetriever.METADATA_KEY_DURATION));
+                  int numFrames =
+                      Integer.parseInt(
+                          metaRetriever.extractMetadata(
+                              MediaMetadataRetriever.METADATA_KEY_VIDEO_FRAME_COUNT));
+                  long frameIntervalMs = duration / numFrames;
+                  for (int i = 0; i < numFrames; ++i) {
+                    Image image = new BitmapImageBuilder(metaRetriever.getFrameAtIndex(i)).build();
+                    ObjectDetectionResult detectionResult =
+                        objectDetector.detectForVideo(image, frameIntervalMs * i);
+                    // Currently only annotates the detection result on the first video frame and
+                    // display it to verify the correctness.
+                    // TODO: Annotates the detection result on every frame, save the
+                    // annotated frames as a video file, and play back the video afterwards.
+                    if (i == 0) {
+                      imageView.setData(image, detectionResult);
+                      runOnUiThread(() -> imageView.update());
+                    }
+                  }
+                }
+              }
+            });
+    Button loadVideoButton = findViewById(R.id.button_load_video);
+    loadVideoButton.setOnClickListener(
+        v -> {
+          createObjectDetector(RunningMode.VIDEO);
+          updateLayout();
+          this.inputSource = InputSource.VIDEO;
+
+          // Reads a video from gallery.
+          Intent pickVideoIntent = new Intent(Intent.ACTION_PICK);
+          pickVideoIntent.setDataAndType(MediaStore.Video.Media.INTERNAL_CONTENT_URI, "video/*");
+          videoGetter.launch(pickVideoIntent);
+        });
+  }
+
+  private void createObjectDetector(RunningMode mode) {
+    if (objectDetector != null) {
+      objectDetector.close();
+    }
+    // Initializes a new MediaPipe ObjectDetector instance
+    ObjectDetectorOptions options =
+        ObjectDetectorOptions.builder()
+            .setBaseOptions(BaseOptions.builder().setModelAssetPath(MODEL_FILE).build())
+            .setScoreThreshold(0.5f)
+            .setMaxResults(5)
+            .setRunningMode(mode)
+            .build();
+    objectDetector = ObjectDetector.createFromOptions(this, options);
+  }
+
+  private void updateLayout() {
+    // Updates the preview layout.
+    FrameLayout frameLayout = findViewById(R.id.preview_display_layout);
+    frameLayout.removeAllViewsInLayout();
+    imageView.setImageDrawable(null);
+    frameLayout.addView(imageView);
+    imageView.setVisibility(View.VISIBLE);
+  }
+
+  private Bitmap downscaleBitmap(Bitmap originalBitmap) {
+    double aspectRatio = (double) originalBitmap.getWidth() / originalBitmap.getHeight();
+    int width = imageView.getWidth();
+    int height = imageView.getHeight();
+    if (((double) imageView.getWidth() / imageView.getHeight()) > aspectRatio) {
+      width = (int) (height * aspectRatio);
+    } else {
+      height = (int) (width / aspectRatio);
+    }
+    return Bitmap.createScaledBitmap(originalBitmap, width, height, false);
+  }
+
+  private Bitmap rotateBitmap(Bitmap inputBitmap, InputStream imageData) throws IOException {
+    int orientation =
+        new ExifInterface(imageData)
+            .getAttributeInt(ExifInterface.TAG_ORIENTATION, ExifInterface.ORIENTATION_NORMAL);
+    if (orientation == ExifInterface.ORIENTATION_NORMAL) {
+      return inputBitmap;
+    }
+    Matrix matrix = new Matrix();
+    switch (orientation) {
+      case ExifInterface.ORIENTATION_ROTATE_90:
+        matrix.postRotate(90);
+        break;
+      case ExifInterface.ORIENTATION_ROTATE_180:
+        matrix.postRotate(180);
+        break;
+      case ExifInterface.ORIENTATION_ROTATE_270:
+        matrix.postRotate(270);
+        break;
+      default:
+        matrix.postRotate(0);
+    }
+    return Bitmap.createBitmap(
+        inputBitmap, 0, 0, inputBitmap.getWidth(), inputBitmap.getHeight(), matrix, true);
+  }
+}
diff --git a/mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/ObjectDetectionResultImageView.java b/mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/ObjectDetectionResultImageView.java
new file mode 100644
index 000000000..94a4a90dc
--- /dev/null
+++ b/mediapipe/tasks/examples/android/objectdetector/src/main/java/com/google/mediapipe/tasks/examples/objectdetector/ObjectDetectionResultImageView.java
@@ -0,0 +1,77 @@
+// Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+//
+// Licensed under the Apache License, Version 2.0 (the "License");
+// you may not use this file except in compliance with the License.
+// You may obtain a copy of the License at
+//
+//      http://www.apache.org/licenses/LICENSE-2.0
+//
+// Unless required by applicable law or agreed to in writing, software
+// distributed under the License is distributed on an "AS IS" BASIS,
+// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+// See the License for the specific language governing permissions and
+// limitations under the License.
+
+package com.google.mediapipe.tasks.examples.objectdetector;
+
+import android.content.Context;
+import android.graphics.Bitmap;
+import android.graphics.Canvas;
+import android.graphics.Color;
+import android.graphics.Matrix;
+import android.graphics.Paint;
+import androidx.appcompat.widget.AppCompatImageView;
+import com.google.mediapipe.framework.image.BitmapExtractor;
+import com.google.mediapipe.framework.image.Image;
+import com.google.mediapipe.tasks.components.containers.Detection;
+import com.google.mediapipe.tasks.vision.objectdetector.ObjectDetectionResult;
+
+/** An ImageView implementation for displaying {@link ObjectDetectionResult}. */
+public class ObjectDetectionResultImageView extends AppCompatImageView {
+  private static final String TAG = "ObjectDetectionResultImageView";
+
+  private static final int BBOX_COLOR = Color.GREEN;
+  private static final int BBOX_THICKNESS = 5; // Pixels
+  private Bitmap latest;
+
+  public ObjectDetectionResultImageView(Context context) {
+    super(context);
+    setScaleType(AppCompatImageView.ScaleType.FIT_CENTER);
+  }
+
+  /**
+   * Sets an {@link Image} and an {@link ObjectDetectionResult} to render.
+   *
+   * @param image an {@link Image} object for annotation.
+   * @param result an {@link ObjectDetectionResult} object that contains the detection result.
+   */
+  public void setData(Image image, ObjectDetectionResult result) {
+    if (image == null || result == null) {
+      return;
+    }
+    latest = BitmapExtractor.extract(image);
+    Canvas canvas = new Canvas(latest);
+    canvas.drawBitmap(latest, new Matrix(), null);
+    for (int i = 0; i < result.detections().size(); ++i) {
+      drawDetectionOnCanvas(result.detections().get(i), canvas);
+    }
+  }
+
+  /** Updates the image view with the latest {@link ObjectDetectionResult}. */
+  public void update() {
+    postInvalidate();
+    if (latest != null) {
+      setImageBitmap(latest);
+    }
+  }
+
+  private void drawDetectionOnCanvas(Detection detection, Canvas canvas) {
+    // TODO: Draws the category and the score per bounding box.
+    // Draws bounding box.
+    Paint bboxPaint = new Paint();
+    bboxPaint.setColor(BBOX_COLOR);
+    bboxPaint.setStyle(Paint.Style.STROKE);
+    bboxPaint.setStrokeWidth(BBOX_THICKNESS);
+    canvas.drawRect(detection.boundingBox(), bboxPaint);
+  }
+}
diff --git a/mediapipe/tasks/examples/android/res/drawable-v24/ic_launcher_foreground.xml b/mediapipe/tasks/examples/android/res/drawable-v24/ic_launcher_foreground.xml
new file mode 100644
index 000000000..c7bd21dbd
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/drawable-v24/ic_launcher_foreground.xml
@@ -0,0 +1,34 @@
+<vector xmlns:android="http://schemas.android.com/apk/res/android"
+    xmlns:aapt="http://schemas.android.com/aapt"
+    android:width="108dp"
+    android:height="108dp"
+    android:viewportHeight="108"
+    android:viewportWidth="108">
+    <path
+        android:fillType="evenOdd"
+        android:pathData="M32,64C32,64 38.39,52.99 44.13,50.95C51.37,48.37 70.14,49.57 70.14,49.57L108.26,87.69L108,109.01L75.97,107.97L32,64Z"
+        android:strokeColor="#00000000"
+        android:strokeWidth="1">
+        <aapt:attr name="android:fillColor">
+            <gradient
+                android:endX="78.5885"
+                android:endY="90.9159"
+                android:startX="48.7653"
+                android:startY="61.0927"
+                android:type="linear">
+                <item
+                    android:color="#44000000"
+                    android:offset="0.0" />
+                <item
+                    android:color="#00000000"
+                    android:offset="1.0" />
+            </gradient>
+        </aapt:attr>
+    </path>
+    <path
+        android:fillColor="#FFFFFF"
+        android:fillType="nonZero"
+        android:pathData="M66.94,46.02L66.94,46.02C72.44,50.07 76,56.61 76,64L32,64C32,56.61 35.56,50.11 40.98,46.06L36.18,41.19C35.45,40.45 35.45,39.3 36.18,38.56C36.91,37.81 38.05,37.81 38.78,38.56L44.25,44.05C47.18,42.57 50.48,41.71 54,41.71C57.48,41.71 60.78,42.57 63.68,44.05L69.11,38.56C69.84,37.81 70.98,37.81 71.71,38.56C72.44,39.3 72.44,40.45 71.71,41.19L66.94,46.02ZM62.94,56.92C64.08,56.92 65,56.01 65,54.88C65,53.76 64.08,52.85 62.94,52.85C61.8,52.85 60.88,53.76 60.88,54.88C60.88,56.01 61.8,56.92 62.94,56.92ZM45.06,56.92C46.2,56.92 47.13,56.01 47.13,54.88C47.13,53.76 46.2,52.85 45.06,52.85C43.92,52.85 43,53.76 43,54.88C43,56.01 43.92,56.92 45.06,56.92Z"
+        android:strokeColor="#00000000"
+        android:strokeWidth="1" />
+</vector>
diff --git a/mediapipe/tasks/examples/android/res/drawable/ic_launcher_background.xml b/mediapipe/tasks/examples/android/res/drawable/ic_launcher_background.xml
new file mode 100644
index 000000000..01f0af0ad
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/drawable/ic_launcher_background.xml
@@ -0,0 +1,74 @@
+<?xml version="1.0" encoding="utf-8"?>
+<vector
+    android:height="108dp"
+    android:width="108dp"
+    android:viewportHeight="108"
+    android:viewportWidth="108"
+    xmlns:android="http://schemas.android.com/apk/res/android">
+    <path android:fillColor="#26A69A"
+          android:pathData="M0,0h108v108h-108z"/>
+    <path android:fillColor="#00000000" android:pathData="M9,0L9,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M19,0L19,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M29,0L29,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M39,0L39,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M49,0L49,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M59,0L59,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M69,0L69,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M79,0L79,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M89,0L89,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M99,0L99,108"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,9L108,9"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,19L108,19"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,29L108,29"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,39L108,39"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,49L108,49"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,59L108,59"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,69L108,69"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,79L108,79"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,89L108,89"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M0,99L108,99"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M19,29L89,29"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M19,39L89,39"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M19,49L89,49"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M19,59L89,59"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M19,69L89,69"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M19,79L89,79"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M29,19L29,89"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M39,19L39,89"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M49,19L49,89"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M59,19L59,89"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M69,19L69,89"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+    <path android:fillColor="#00000000" android:pathData="M79,19L79,89"
+          android:strokeColor="#33FFFFFF" android:strokeWidth="0.8"/>
+</vector>
diff --git a/mediapipe/tasks/examples/android/res/layout/activity_main.xml b/mediapipe/tasks/examples/android/res/layout/activity_main.xml
new file mode 100644
index 000000000..834e9a3e6
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/layout/activity_main.xml
@@ -0,0 +1,40 @@
+<?xml version="1.0" encoding="utf-8"?>
+<LinearLayout
+    xmlns:android="http://schemas.android.com/apk/res/android"
+    android:layout_width="match_parent"
+    android:layout_height="match_parent"
+    android:orientation="vertical">
+    <LinearLayout
+        android:id="@+id/buttons"
+        android:layout_width="match_parent"
+        android:layout_height="wrap_content"
+        style="?android:attr/buttonBarStyle" android:gravity="center"
+        android:orientation="horizontal">
+        <Button
+            android:id="@+id/button_load_picture"
+            android:layout_width="wrap_content"
+            style="?android:attr/buttonBarButtonStyle" android:layout_height="wrap_content"
+            android:text="@string/load_picture" />
+        <Button
+            android:id="@+id/button_load_video"
+            android:layout_width="wrap_content"
+            style="?android:attr/buttonBarButtonStyle" android:layout_height="wrap_content"
+            android:text="@string/load_video" />
+        <Button
+            android:id="@+id/button_start_camera"
+            android:layout_width="wrap_content"
+            style="?android:attr/buttonBarButtonStyle" android:layout_height="wrap_content"
+            android:text="@string/start_camera" />
+    </LinearLayout>
+    <FrameLayout
+        android:id="@+id/preview_display_layout"
+        android:layout_width="match_parent"
+        android:layout_height="match_parent">
+        <TextView
+            android:id="@+id/no_view"
+            android:layout_width="match_parent"
+            android:layout_height="wrap_content"
+            android:gravity="center"
+            android:text="@string/instruction" />
+    </FrameLayout>
+</LinearLayout>
diff --git a/mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher.xml b/mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher.xml
new file mode 100644
index 000000000..d372a4fca
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher.xml
@@ -0,0 +1,5 @@
+<?xml version="1.0" encoding="utf-8"?>
+<adaptive-icon xmlns:android="http://schemas.android.com/apk/res/android">
+    <background android:drawable="@drawable/ic_launcher_background"/>
+    <foreground android:drawable="@mipmap/ic_launcher_foreground"/>
+</adaptive-icon>
diff --git a/mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher_round.xml b/mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher_round.xml
new file mode 100644
index 000000000..d372a4fca
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/mipmap-anydpi-v26/ic_launcher_round.xml
@@ -0,0 +1,5 @@
+<?xml version="1.0" encoding="utf-8"?>
+<adaptive-icon xmlns:android="http://schemas.android.com/apk/res/android">
+    <background android:drawable="@drawable/ic_launcher_background"/>
+    <foreground android:drawable="@mipmap/ic_launcher_foreground"/>
+</adaptive-icon>
diff --git a/mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher.png b/mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher.png
new file mode 100644
index 0000000000000000000000000000000000000000..920cbf3f197f79b0e6a6013d17b855e916be7d66
GIT binary patch
literal 1354
zcmV-Q1-1H#P)<h;3K|Lk000e1NJLTq002k;002k`1^@s6RqeA!000FMNkl<Zc-rlp
zYitx%6vt<`Z6X*Gj8!Y3NDOKMY&4Cjm0&AY5{#Of7!Cd4Cu3s#Bz_>SX=nl{G?cQ?
zeRQ$4u-(44TamU%t9GIF5ut4;=~_%(-Zn8nq1*b256Al-cBa$WrTb{n&fR;HlbOt%
z`#8V7Gjq<JGb>4=h$4z8q8Pgnk%>trXJVQ!IA)H2c=Nn9;QBrEdQp2|`_qjbdso)C
z?^)^nw!%=t#{4WBJ8Xb0bB3f;91DlDUp?9rkQ@!flN2$V{ump}P6?9s`Yby7;|aOZ
zs)pNK3A3r7siT*fXx7`D2%t=CtImwPp~B&821$;_+tU6+w5`4C9j(pD05ZMeJGPr!
zKz0VGs-9kQ*4bcdDuB$ZYMPzg0)kEUdRno^Q43pB0hIZKv);w^3}S$meYyW@*qRz3
zm#3P4K(@Nn0XYx5`3JP`kN_kA2|)Y+DrozjJi$>q9=b%<p9konn${?twN0JmJTpSw
zq03Z$`Yf$Dd?ZQ-$8Zih(6z8SdL6WlKG2tEcvS9tK?A}7uZ-WK$*F&2nV6iS=8->X
zQFXJb^NhEhf`4D5a6|{&lb8OYhim<+4)$?OtAjeItLUIjw2eNL{?K>Akp2%q_!X7G
zDl&F+g4VZnt2(8>4624Jax*+dC0%{0&SticV_F^5L0v^>y}w&-Ti+jM?}7!?bKz>#
zfXBm=RNQ$&)u|pDiPC{=={>FLd>r^K3L?Zd7*uu8Cfe5PU<?ZuP~c)LKvVSgv1vf=
zy8!xB18AMUi>_)eRmcqn<Gtg(st(#5*Ud5r#=t$G3}+*4?dzw|wQ(8DWd_SOFieZx
z%>s}jW(H|ZV<)}Y(L*ozzN0Kx(~R?J4=A1pKmyS0fF7*z(T6?1(gF75!V_*;-8TY|
zqVvGP7HT~gqFWIs5YAv9a~zjyAf+{+;?4kFpP1?60MeG;Q^p4L!}%*w1IA<eEoN{s
zfIc~SDyEK$#bkyCgy2%|0mGNU<2o5Y*pH=yx`qaX2pmBs;wXTEHw=dlK8*mRiNq0X
zBA!CXkfDJi03{#-kN_0_5-SmxSQ0nspTtdCJm%-qQG~M>%Xm14+yhD^)IhkPU1*0O
ztSHt%Olv?0I1zX%kPwg}FhzhW07V2#7g)l;QU;bZu(Y8)pZ0*_i2x)3%?=0)@K}h)
zf;?ORv(A}JYe0B)V;cy{uX5l5z)wQNul2ao8W3Lo-O#-eg~yGsgnK}Vyn4p#W4#W%
zZ^UlXzEW&J)fzyCeD#dPftbk){`koYn>EP`C5a?2wB03nVR%%ezZf?CMNOa@o;`ue
zq9jnAPdtHYZM+1kpaBx7L<*h&BmfCO0?>a6$dLx$YJ4$=Z#ABBH}E@yBA-J&S#D2#
z2EFWQ+sFF{b$JzbcgldadVTLtPe(6DaJ+Xu$*ioU56UY)g021i{aML=)SJQNv2_Q%
zgP4kPgLY2Dap(S%SLqyLHZF&)N&nG6OG^s?EqQIzrh*r%z8aGD){u0-LozHQ3%jgf
zOJ0TZ?1qgSH^9cSva(Dml~Pz(cpuv=W%As+pOsplFWcpPymXg8$GX#(^ZAa}8DiD#
zTVtF{MCV?#DIeeITh7J~8(<4;f^GTv`HSZA=L@FX+}y=%o5N%klhsVt7+h9IY<bjT
zvD^<E=lmZR3{n=GO-tA|n>G4~kp(tHY{`=DK{0V6qKG1jD26Ei0;e{GZ^QhPyZ`_I
M07*qoM6N<$f+3TIaR2}S

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher_foreground.png b/mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher_foreground.png
new file mode 100644
index 0000000000000000000000000000000000000000..7dfa35f78604c7e583f43af36ec3005bc19561f9
GIT binary patch
literal 2257
zcmZ`*2{_c<8lM^4kYz;3lKHd~nUsm-4n|oAWmj{{#1~m&2-&6SFEgktW4khvWyX@-
zm9$KwED6J88H$i)COf6_{q=p%y|;Ux`#fiT&vVZEzURE>_j?m;ai=B36vZGAh=jQr
z_6*q5wklj0Y@aX6DM27`syWukK6H?mYw5GYL7^3zD5^wbVc9otSGvb)Ilqc)2*^Jq
zC}U!5rOmKT5n^1}hPs0gC&{&&goTF5R%+I;y}c~+&D`1wq(l$eUXE3>9dx_!a^c5*
zuR}ZoT%%5~=lb`1n-80pI&;26h44G{rcUU1Z8wJh7LWL)k$og?EC<610&u7r3>y!T
zhN>CI01)ZlFxVJ407pn8VW3<PkP&s~VEX4i%^jiJ)Mrf``@Mq`tvd#15b4CnIw9r3
zZ}MQY0GzRDNQKMQ2ZOS*9vu%fTgF6#j4rbD)+RXDjYW>2kgF%JyBk&%3jowmJCEjj
z_vFQx8d4&i8{usbq~||s>pNa|ZiCXE1bb^IBo_t8!pGYbt{{R{X?c+CyZmR<r@gX_
z*w&~79BBoSX@OXg=#^14$C1+L_Ho6A8o}}@ALD~~vK~1(g_79d`*+`8auE<O5wL6i
zvIp(r_~wiVjg)TYSH;*`rV#HAUvkkCz={Jg090B)4e5?yf(|PnP@u!Ye~MA=Emgex
zE{>6HJ_a1WD1*_9j_llHpZV%4PUI5PjgIGwGM?WD5<S~*;;JWu=$Rq+d#to-l2Lk7
zVImrxgTr;VA8E<YRxc_p1!}E^CQUSR3vo*9bc!YqH^e@LA)|oA<i4ycYa5+!3tGF@
zf4om|8G7f9ucAun?I*I2t`6^#*?nU9-6j1d2J-Fi%LC8yTM1)1(~aYEyAP0(99kI!
zRTUFodp9VGgj7elWH6JLNt3#x3e{_=nS2qcNCNxXT9%jFS~!ck+8Z^c6NRct9)LQK
zm0SlS2*qV~FW^Ae0&(uzUS;01{L`L4^3}VJ7dn-Tuu`1RN5fgp!Bqq5q}7QXgedbg
zBEty9o;(o!5E_uvA9>GuabnCe<2<3GgPZ7~NeEiz>YP0nHMxTR-nHsnV^dE!_#|{d
zpEM{5prCf>g}Q_*d>(PrtD?9%36;|1+W9Ul)QXG_j8>{_xGS--EjdjqE3oyll8KZ@
zjDidzQR|y;Zo}&Ej&@C!Qw!R;HDz>QWr(WazZtk^-Z1z<hH^;7f}d)ugM+UK78Ajm
z$!e6Iuhi*L<ySzk7;#nXm)50R$D#V+8DevnASe-HqN5Fc%14%><34vuX4+V6&ihCq
zVeTWyHsYECwdM0j2$#>t^dBnIpB~zU5oFq8Qj6qOk7~?jcP!A+)<e0`Rxc=SlT8mT
zC%N2bFDPKB<i}Ezq0)%|D!Ya3dS4nEp5X;?B3vB__|X730va%eVFRzxwb7KF{N6nr
zRo4?a<tm^SGb7pf>^EP!giB}n$9f|trD1BwNr|zHv=zr8|HaHFLO=`^4JaBt8u(#!
z?LbJEGYp7PFa#?64hKN0Dfy*_7%OQ&ol|3-!-_0`*+nbai1TC+X9a2Aw|r%a%;BJw
zcbsp|-AF&X+*0Nb!(zyj@m>x>J8DGsS`!7Ezs^IX5hirsRSh2wW4du8Da=m_RFGBM
z_PmtjSybojky~2a`-wDTcJ$_`zZ+=q+p=znJ_Y*Y-fP{v!Kt&tFYfnlRvx)=YX|1r
z=s*4I_#ypT$^((arDrD9jkPZxo32S9S@|YSv-p-TPYFM_UnQ-J@ks0XJI4f=X8TVn
z7uVo3uI|0J3xX<@sq<9seOgKt76z=yLRbqSn1#?$-%(vs8x%i;|MMEuFIAPH?P62i
z#6eRY8>QZuxpS-Rt{pgeg|Y7rKmXLNK#<NE<ZQiF%yIF8bfKN!B@L`A`;MrcjTo5{
z0bNxDHn}>x0_AW$Fbh|+*Pp+KZ(UjJx`8kj1IEpN7P2h1Ira46sc!gf*27wSk|$5F
zY+3B*8J2X#98PMOrVEXe4OK&)5~JbS?^Lx%7=@OP9VuW$b|_&f)SRYZjyfq_8wx&y
zAhww~q)Q@g43`Rk0kk9>@Ot6;tb)3wZHYP(Hl{OjSufJkZ-z&tESQ?+28*YadfzGP
zM%i1IZ4R)y4Y^5$dLiUGlx|&(fcrVjIS`@r176A>Zemdzo7noKnTK4@51;&1(w(Y(
z-sqeh;hu|_u3C9i%-i~04!se)sOSql!bB%}vPG%Ecq+cpFYTMd{_K;4;-{nge*Sn6
z3#OcJxEFU;_?&<P0qT{1k6`FlK3LxP7rxk4Xe#k!xXo|k?nlwB*}_)keTFrMTL=Sh
zd*s#&&tw-E6va?Q$bL?ztOL()!P;UbTYpY3vtZJyOC^K(w`;f8eakUl$%eD5FE7tE
z^8Ug!gX?kp`7~{BI20A1=fy=&?~7XcB3Vu2Gs3#Y@g40mSuW0aLVs@ar_@q0cyr^0
z#p6zSZ*@x>9my+=&fD5;s3qx%#uQ~gdnLBq$<z{GyC2E|YDu~O597C>e@tGH2{M|Q
z-}5bQVD=Tcng2CZ$vwXZeQ_kS-DL(dr}!!>jwUXO1Gb6j;I`<b=Dpc5a~zOHFADpp
z-;Z<*TD$u(zTumzjbTX=8u&_z<l7l1jE|e9t%RMIAV|^JYo8J_YwOAJt?V{R;H4f(
zTF}G?q&7>|cw(RUUKsO(#<*It{M7q|dcVx)4cqrgGCzc$;>ym7mfcS&&lUYlJJ@!7
z3NLlX#=5~m7T6tYpO>a@+m~IDzn4NBn=b>Ym;oh4d=)(_s%DY*3zGu~q#6eMZ#Lgz
ic#!D-CmZ-ki-ywJbH*KvQ@6lW1!8WB!<HLgi2DPy6%q{q

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher_round.png b/mediapipe/tasks/examples/android/res/mipmap-hdpi/ic_launcher_round.png
new file mode 100644
index 0000000000000000000000000000000000000000..20dfc6c136d063ce49839b3f0201e424b428f6a6
GIT binary patch
literal 3246
zcmV;f3{mrmP)<h;3K|Lk000e1NJLTq002k;002k`1^@s6RqeA!000biNkl<Zc-rM$
z3s6+o89wZ;4{!vvzEFI|SGw{@NsMSr*F03T#;GG!N1G^GW7<S(jY-7UCc27>qQJ5c
z@Wo?+MM3dRG$z$l?KIO)Gm~_*pi!dH`hYc3I}vWrckY9`Pj+E<*)@0O=Pq~8{m=h@
z_doyn|8xGkgtXzoaa@o<3J1i#CqI=#P*_Pxg|i#?N%uT)fC-iW<rEIS?_f4_-S5`d
z_p5#M=qV6B;~Kte3D1By5=<tg=FAzD{Mx9pQ$LG7aAJH+>52DZN>Ap~^A|Ct$E)c1
z4n04NF&+OOBYftsRq<_TThs9sxEDO+JkbaG3LIp#6cqU|$T9XJSw#y-W^n~E6x|}3
zMy^!^9)_Yrj3;<EnN<+$+iwp%MXk2U<g<)%ly*~P?pZ2X*g)wqPsYbYU(mo`Gd5Zh
zy~e&XK9;p_v;ylM>jQmB4~WX(cNQgoa)u%eF&ICgXZe#Cm27CW1xWydryMNSr^c>K
zZ3hf}TB6nDm<CY+OD%!1^VB^{jY1!Ct14x(Sj*@0BRjiY({L(YN(K9`mC)V<#t~4B
z0!WXTO!j2;DLC1tFe=ayNze@Ak3Iy&KFMV;7>_a;eF~D;k)e1beNtDOgz|)c@f1;6
znt-}W(hDMe1Cr?Y(hDatGxis&Y2KP1k?ngXdDnT1=<5q+G&y1`2kDn2$fz&m$hMrB
z-m#=6a3%y6IMyH)aAZr?WY2;mn&i&Bk<zdVC~&MmO3!;lw5c(k8E#?>kya4OoL#_z
zV+GRAz2A}#Hg&bou~h)+g{iz5hpeK*uVZ!z*FLM53v$v9-=PiKDlH>!8&bA_r^*>m
zQHO20$EjvPEeA+6zl_2qu7DF&e3VQ1+gIG_E7v*W*Z<}cj(=`fP$)3B@+@ciri!bq
zzQ%3(>p!`ng%x&r4lSr)-=kd#Z9lj29J}A%gF$~qwBOU!*O?rU$DzFXq8~|mK1Vj{
z-)-4Y6O+BvJp(Dj`lcMnF*p~m-{i~>9&!(U;OS;<9k=My^A;e%8j`;FlDm1oPHI>8
z;0Nw6*KctB_LW=k#Q(UC?<Jn}KHM+2I`Bt7=ud11IUtW?kf<);k-DdztXMzPLT9%C
z(ifC8XFU)QPfQMuNEd2ua)a^@No~XTmT@)rYHiwqR-XIPf-nBJ*X}!dKlq64!2bdT
z!@3>hfm{xU8+5|dpKK^&L(OG?q@lI<MPW*Jnp45#F|6z<gC{;_|4FIs_=BhHwgX|v
zb;vGe?Qq7Ba@1B+{)xmB{N)xYC&*=QNXio1Mv}IVBWqJ8x<1rIr<GAy>B_7XOa=Jt
zYK>h1Z`GQ)u?Ly}diBt0he{v#S2g1$4n_UnzqZ@%J{@u+i;hS<!5{tDYYVx8odG23
zwmIulmmIdVbeXnqlx*3e(boWhUM#ER?$kZ7X?OVY6$`6q9Y9n!=(%GrM^+`j*sL#-
z){xWlb6t&jyaf=fX!gmo+_`HvnBTelpS#?iuP#xabI_vyRszK92w$YzneTL%Njn4^
z(|XXG?ld=Zc^e>*f`T+4?;tmN{|PRp<QNy4Q)<({2LR${v~4+e$()4|4ueN*qX<)b
zqL8<^w0z-BfLx*01H=;929lP?k=4n6uoy|!n{&)jrR5!TC#85I01|cGkamn>ps?9S
z5_c=r+U?z_<=k|2czGKj_@Ul=%9z8Bq(NV&k4@eJh#zKKvu}~<bHnAB$gPK{ca@BH
zW#;k)K&q@FZpmjCxvO{Y3laB2?nKo!29)Ok5@I6EW`A5e1z4Ps4>QsKMD5wX!rhGB
zTL4X{IL+O*Qryfy`j%Q!i5CE(IwL<@mAn?NM)Jur3sOYpm*@KfKp0wC|M<pkJ2NGD
zj^#0c#OYp}S^_Mj;tE-bi89()an@ac5VRrNc9b_^igp(uobkV-w&3FxfN<8OpQl)~
zmv9lCkw&BG7+rk$vO80}__VPkA<w!+LXd|zCLjQU&T!PpnN_5F_s)`|76U{>US7}W
zl4G~rnZe}iHdvaF3x_me_Q^l9U@t%{hzDrP$G7`VpA#Vk&murY<>dooijUT{1R%D4
zlUL4`D3FCC(Ztkjy;>L$0By@P_ep$fxD7zXWo5%+N{-gMJIz}YmMR=Aqxv^MLgKn@
zSAE37g;6#D<(o{yJpqt23KW3)CMHJP0F=3V_u#0KBj%P4vB0+=DQSJsd(Qly%K>7u
zSdvZ`9$j+uR*QYmz9+;7ZOOjfWzwX9@>z@l)SXUk(S;RmPxEXRD|9rAoei8>Oq%BP
z)g&~un*~4uNS&hc%g(tQC_XW9kf`&AsTV00p%yjTLRcNF&MhwdNk`I1HIg>%2No7}
z+(Kg3{^=dLtG$_vA%`+MP#U?H=(E}cASA0;G+UCiQAjmxl2z5<B;6EuowGXtAqz(~
zZoP+wtR30AoTpC#g5|J{#N8a3o%kDIVZGdev8L!U=A{7?#T(vQFau#d2kSa@0!Ucz
z!MYFDf7%cr-a&8Fe@~*LhY8ud<tq;YVqs*n;jp)sT$?P(#DGF5fE7dQaKQINOz=M|
zb(Iu@E*h;?<!I49guM6B5^pWqyEANk?1LKEYU0ATKL7@?_KK?nlWIqPG5Tqew(qt#
zwtf5(AaQtY%D6++>KKNBqc@7wT5WrJ>PFV@$oIy^p-9-@!af&50K`d(wj5%Zy)p_o
zvV2`B(>;M<;B@Obgv^}(D|VBIx1Pmn4D66%rxbA#f=NW&azErdfC~8pH6@25@zZB8
z-R=1wiz7ZB?`ZWV8`DpFx_ALW<>AX$ZQ3C%fx~VCfOy^B-*A@d94_eIY;^|^>`Z3P
zOCafa&03K=0HG@5zMYf=gap^OfuhWMrY8KB>AYx>i<Pe$O-K5-4_T9Ph)sV^3aH$v
z4?<m!Bc)5M>Vr&9>Vr^c)av>ml;OdREL(ew>B{Tu>g$3DN``3c3onqAT&HV9PAkl=
z+g6x;*UAdB8haIHHEvdzi8Gfj!%Wn%<C(4+jhl6fnt2_WFIxPrmQ`UHjE~)|QT$yS
z)+k<ZRHMk!3aZQYT^xCR@g}BYTtaY5>Rf}!@Zs407)(|tpJe6;vuviy^;EShRj%hO
zt6U2z9Ii8oRjwvh<$Bg>l`Hx^b;U`QE5OMVht`_YRJW%D9b5P&HE(uZNtZ6sM3-LW
zN<pVz*GY@Auh~*f!op&iZn3egd?g`vY6$(?i;NjJg(Mryd}&^QzKQ_Z*YA8lo*grd
z>BQ^if#dn-L+RfL<!iGN$=2+KzynDDXOq4`F?sr%RF{5&P9FP_f*=wT(}kV~DrU@k
zhiuN`L4r?W<JwPw5knLaoN2R`K?f2U+0~Qah=+poXN(6cr_NYNHtL04SU&|xbjYl7
zY^PZ<dD@$7!ii9KW4d_hP>>NL@XH5-l%vPKM%JZN3(<B1i!J;IoCYQn>U~yku2znU
z`!!@1WcT`8&<dhvPhkKF33-kz*H*HaLL71k;oHvFG%^_=3uIch>J$kM9>ru;d)IPo
zebV4qyiMGf(n!VYi*+;%sAW&AO@NXC0~t1?)hgyLNT#wyG8uVUeg3O4L8MQg5YoLn
ziXMll)X@{k(p5+3EXTk42w@FuDYL<=L&+e_Q1ChEd&$b9DzzFf@D5|LLB=)!M}DYC
zUpBV;W6e$Z%H;WE`MS^9Ob7qOfGTQe0;@Ha1xpcrnu75ZECzkiK9w&|SpXTBEWAv9
z^J6ExHTCS-i_H{-L@6~d&n0iaZ=!+O_iQ^wvZ^9J7cvMxSc0|hi~`=t%ZXhp5WoN#
zA4~Ux7kGj<?L+y?m#|zB#rlQ}eqc3@4kZ~fBvcr707|XuJ7BzG*1R<|_c=*EOs!$-
zFIXTKS1cva91u$p3h<sOikmp{(T*DOr?n@*6TAhB8O-`b-;f~y4kb$hdk*TwqE<ps
zu@$N65k5}&>ePjbH{RMx-dTBoB<ar6P~j5MrQIOOnRm&StU5;ctV_Q^)@{GU+AK{v
zpm^i=TX8RV2-Y&34JGu!C!7KFBTMp-V91aTVKj2>#1Fp+mO&FT0<&MccI}>_^jydA
z@Nr=yMo#D&`Aj?rpK*<T_bl3=t#A+SjQ}t31aDtX-ChSvtyZ@mJ9cd6xVX4(QBhGn
zd-Ukho2>=&!1ZVG3QzroYvE{vwzvoPg2$5tigU+K71d}os@T}rcJcA?9Z-r4!e?C5
g@XIZVwsg4v2U8oiR~-LY^Z)<=07*qoM6N<$g59DoWB>pF

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher.png b/mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher.png
new file mode 100644
index 0000000000000000000000000000000000000000..3813caa21c6cb224a8da4a76c31cfeb1f93b9833
GIT binary patch
literal 959
zcmV;w13>(VP)<h;3K|Lk000e1NJLTq001xm001xu1^@s6R|5Hm000AqNkl<Zc-rk*
zUr19?9Nyd<nb3<t6eScDp;-?m1hy9~VVXgZ&_jWuMm>d;34&B$e+bL|;F2^UtSQ1G
z|0OKl)VXOkQ=`^XC0HTa=A89VD5vus@0GjZolDSrZ|uUy*}i+v`F-EL_x#Q|_o!5n
zL&gUoL_?vK7@Cy`j{;<LBg5AtypaBmqll;2CJ_)jZYgac#nW>AJR!!mv2PRkUHzWS
z_8yNx0w8X_bzq|9%U2^bEo*{5`Dbpf`h4FI(N#ARmC+<QpyRBiw}whniyGiNi{J0x
z*!p>VazXsF-xTwzs;ANbv6%)F(cXPBb<J$vEYuooK#S7tcJIpTbWAh%tPzQy0I#94
z&S$t%ks#E%qyf9#zB{Yc=9Qy?l#;T<unc6jSY`PH4qPnV8!;e+`eB0zZ~8~RVQmc{
zL!GkN$-TE9NcxLTqTvKYQ~pFp52+k_PmZ<QNEFAT#DDNOAI+s_)I)u=gZ7q8+pA#!
z!DfH{Vwm*#Npe#M<Bz}WCLaGR<ELi)B#nv#j}rgE<2<edk9w%jw}p0B-2m=p)v$Xp
zC{WntWaqBG9u44&?JmX_^$Z5^R|iHIUqI^#d9)i9{5q|M0prNWfWT+0{p{S$!FK`t
zCC<S9#&+^)%FDPh&ont?wlE&`gx8~8SsGANB$(RBX-gm3N6qkkMllA&pcn(FbELV8
zT<#wxhF0rhwo!}$J$>iLy*|c$pPpeRBsFlK=51l%+)hVG0~dPV1n}hpqtY1wFeveP
z9oYpPG?W39LLS#iY5+knI3d3yMXEVoF$TmCc><9>u)-{q$7La87a@=ABTEBf;{ftI
z@;o@?eLXklQH%kMurPrU7b7n(MllA&pcn&~Wn#98StmNST4@F_&&Pb9&rV3OvNRC9
zA}iolWLRkld5ja_i-Qbc<p#b`iWQmS*TnzDR|&;`7n1*%IVFGGbvm7UvSj)Ml8efc
zLK$G224=H)Ywjz@kA((gvcZ8qa=X@(a;|6xG}>%7ZKy4#bv50NCLIj$oeQND;A20i
zC%V!b&#AVUaQ4VFsi~=nCm+{0Yb$EJsylTvs)|}4E7uZy%v+DVx{~siO-V_3eKg_j
zRK`$j->287?=&3Ej!(;w1{sFzEhz^NL(8&vKiEW0PL7WLjHTEhF=C-*MH}zvEH0NT
hT1tp^*vKK1<1c=9ZjnA(;xqsN002ovPDHLkV1oC5!>|AV

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher_foreground.png b/mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher_foreground.png
new file mode 100644
index 0000000000000000000000000000000000000000..d492327788b3137fb5a42decebcb7c796d082b91
GIT binary patch
literal 900
zcmV-~1AF|5P)<h;3K|Lk000e1NJLTq003+N003+V1^@s6k7ifR0009^Nkl<Zc-rlq
z&uSA<6o-?IxGlPM-`B8@;I3=m!bfNpq9CG+DhLS{6qJA{)P<Bv0xG0Itb!{S22w~c
zX;IPZ{8|dh%}kuOnd!OQ?+0fw$(hT{{buetbLY>gR1iW4A%qY@2$_;Yr3zpGLjXen
z0~i7r0vN#P!bpJq{`1$hdVB8Sr?2i#_j44ceQhWEtbGp#m^SwI&aS)i=$WhBZA59>
z_gkIaQ-LcL*x=_cclG(}XhHe5lJtf8#3V2sO)J2fF&)4sInzo5rXc`8>82223Yd<d
zm}6F+n3ne}D+QP`Bf!izWz1%PDV+t%d{ZBp0j6sZpv*V*5x@WjFo0nI0~o*nhD89Z
zn%L6%6L<0cMlOH($-jEeZAS_0+KV@?)9bs@_`vP`7`gTC7k4S?ury!&uK(pHf32tW
z&!g=qfvH1A`v-39+4i^Z?!vvtQ@&c(@4m0rPwUts+l~^Lh8!Km<CIR7t-jov^4D^f
z?@rQAwWPMA1*SNmp(m@OYVW#J{%gr{map1bni%SeY&%+DmmfUM>?GOE_aCSHwLH&X
z?M)^QibYwzH?}{Vx@~|J08FL?BF%3&+IP$KCcvx#Q`YAv5`zTR_06|4+X65Hz>c#&
zB@~KL8U~je&D3oHm=R!k4PXER7{DTmVYTBJRwsr9l_d_z62rR=Q)=&LivVM$Bfh!B
zHE+J}B%Nw0p^0|39$<mp4+B_Wcbowh*fSjf3+$;WfB_6(42dKi03)pj7#Ge=P0c=v
z2C%>;e>Gv7<*V2?nXmxN05IJW(XEj*UnMG5?`B5}Os1QULl2q~6hEs3m=R!yXIi~^
zI6NxGt-jit$tD423|MFn_y7!G00UT&F|2VE!}JJ&F0P4T#<Ro$Sz;j9VM%s4<T{L*
z4uBQ(m|JauugspMZD#8i@Zxx<z?afHyp+yoANa5eA6DVhMSQx5kAndWU;qOczyJm?
zfB_7%2#iu5ftB)T_4+{hkF_lo%b^@qVCASJ6*U7+x~UjmUX=!nl8S+qRFo)M8c@;F
z$umtuK<s6K6(Y=+1@^(UBP_h#3a#)qI-39nFa$6JFn}R|A%FplE({D9LI@#*5JCu{
ah3X$qhZDu;^5E700000<MNUMnLSTY5sh2_k

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher_round.png b/mediapipe/tasks/examples/android/res/mipmap-mdpi/ic_launcher_round.png
new file mode 100644
index 0000000000000000000000000000000000000000..18f9ac5e7a3beb9cb42e929dd6c0351680fa8ab2
GIT binary patch
literal 1955
zcmV;U2VD4xP)<h;3K|Lk000e1NJLTq001xm001xu1^@s6R|5Hm000MSNkl<Zc-qxk
zd2Ccg7@uytEtgP|2CRZ9Rut@CghYiu^gkjo5{Mzh1ERrTi&QF&(jdo%7U<R<^muzf
z?{>RO1&vCGD6~|KmP77Jps5t3(xMTtm_Teh&Trna^S0YP+P7`k<m-EVGxL4lZ|0lt
znD+=7;g3=p#VDGQ8P^c{Yke`6jMdLVsi~RL7=LZew{A`PzVpw-mhSFZAl&1b-@c3o
zc#Wdw789{F>lNSQd79Wf?-0AkL2M1jh~#c3lB<U}JhwUF9?#Im`>?(kqvOkXf_FrJ
zPN!MBml`J#yL%<ExeqhNdKu9$sh5uhkJ{Raeq|W;gRM3;Mr-#DcPfz@F8e4nK&1Nq
zkO+OnvET)s>Yl)A7`8uD#rf>GpJxJFy~17nq^#jTehqsC{EU?tZ^-Z`d&J7y8ZwBz
zp-&TTbM+0wY=#2UrvZ^=WQ-hyPI<RY6UJxF#{`U_0fo0Br_M2a5TZA*xm?<;%l1nn
zN?eAr8l^Q2OoOQKK!}@1a@hwl9TgHLtEk$kfe2}1t3F<GtrzTFr+lo#L!<l%j#Ao&
zQY^f2Tqb}}Qc$r$r&~}zU*dXB+pvP!JX!Z8wV%C2Pu%FH#i!e7V)f^mz!^1PQ0KWz
z^!SaNwDiocbSC>>ZNojDp$&az2yNg2Uf>DdkfApxz90HXj<hgvoDl6^s&5rGti})m
zdhh+q=}zxGn%#PqM%&$d_6km2pm*6bwLS0T1#KMe)#vDgz8HhC;BiM7=U_5)E1~L8
zRB?i7W*VgHVAH94Pgp|D+l+JWqrY_ZXo5iBc3kJr<I6vzO`SJ%ZO1yhXq?kS%}mjz
ztNOMhzx__9urU}59%370;JMB_!Ia9FDOsg;eoaYBrkr0-V77nR>zlnZF@DcJdioDN
zfbW<hDgY=fwzXXA;%%_+oYr&NEaSTK_&L4;Xa^5*9ArS&;5dO*sJuebEt7RRs~I7`
zDXv;6%sw>&0LCU%9-x+=&(MSIoiv9jKaKsbu4-riIQ5g#I!dy4uh!|T%gIaY4*RqH
zkpck0-eFR>ie?!y09B_v=Mi-;aNV^cB+oIM<s!Y2SOL|A!tO<P8QuyoodCYLx|e*o
zB}FZc@JDt_g%zuXO+$?YFum#kwVpW7`d?=r=%VI+!!#k5C%Ff>8CmkzDgeyh6)39Q
z>+AQ8+Qj*Si<RBzMH=h$gb#qH<1R{aN~;l_#S%7(?5_SH5C8}}+10HR_D#n%_S_Q?
z08GZ$#Wr2_y3w!!;A*x1#B`1kiYr!KmDaQe0sulZ6$2SO3pQmS03h7^#)$Zl0LXWC
zq-AAI)d1L!9(|mY)^`U3fRq4H6@dx9Jwb7YqYj{;q9?1UD9M`w9XgcEBHiD?0H|d4
zNy<hYKvC6gi_<wr15mtw|Llhiz{bMD6b(RbZS51JthVQ&F`>fpTj|-^$r^w+($c0r
zRGT=@*|BKZvT0&I7spWjI9xXhmk8IWbCt|>stKFn0CtqMGMy)=xlj!Esk!#5;RZ4U
zB?KoiSRr^tDZmW{fIzl6ub$~_=DI7XL=@R#T_>ax1K|KrD=UCNx?nwdo~IAQSO5cZ
zDp5<x`c0WiBo<PMRjKosC%QYFWG-03W(WXDN#+VF{voffc!8(X&HMyUOQ$EX&CxWR
zbUIkdYzP2VCreQa({+lj6lhlWLFU$UZexDTC>Q_~DJWJ@w7@$hR2peQF4=ranE9k;
z!hB+MNBdfNmYh73rPWu$Eh%9#C=<o95plw}O|0m$v#gyYB_$0iDTP9L9Lcn7;%6~H
z9AZ%qyFcQi2!b6VHcZMlM!?0!hf~yRv%g^**U|6!koTw}k?gXc8mnU!`%uR!+X_!J
z-DYqNmoJYFbSf7|UVCdPs}COdn$j2?Z1u^Wr~Cm~vSK;cZB4KjATo$}^8Ush+~;EX
zq%k;Jl-=e4tAQz5lVRmL1wC3+4^ZVLdW?LO?cz14-QDjCBC;pcT={-Jff<`UTo>pR
z@>?5uKVYe5k*pnGDJvl_Y2=UBWwU|IkGCIWIy}jB8Pe_pzu<_2b0wME8c0cv3_^Y_
z6oD5UF8KwAd~~AD{8L;9izQ}wZ)xO#0Fh)$-C$)I>wWEb!!RJ4;xe!Nz4uLP*5^Uy
z$dnh&DJ3PA+u_SgUniD=Q+zc<Ad;&;%-2q`V-OVw&#jF2q9t!~S&$jF*#z9!7J$G>
zz+BV2A8aQ%C0F=T%RrPp;tJt4tCEzd`4V_AUZ&OQ*^KwITn1!C<TC`>2g9TuB@^Q3
znO3E3BA?{8kWcsA;WLN`?mb1*lXsI+9#t`Bv$ciuV7xfbko>$6t04;+&uBWMXUP18
zubAFmyM=9b4~v+sWS7)V@*Fowe)(UVaGz^y=WVPH@5{!pv2zq2<2lcWe`?_WV-J9t
ztSEwwH#Z?6asIpo3toD0@!~}w+~XPAls?myG3L?#*u&4~HYhCG+S+0p8yjO4Kv*4o
p|L}-H6P0IX^g$(sF-ClD^FN+2BkgDNGH?I@002ovPDHLkV1nt`tv>(&

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher.png b/mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher.png
new file mode 100644
index 0000000000000000000000000000000000000000..b79cc77a0a1e54f90787a8231da0b669689b9ad0
GIT binary patch
literal 1971
zcmV;k2Tb^hP)<h;3K|Lk000e1NJLTq003YB003YJ1^@s6;+S_h000MiNkl<Zc-rlq
zZA?>F7{~j90*af{xlPa^%hX6z+;F}Q#s}huEz@jBSfbfJZ8L1i5|^1cXU1HNF|seb
zm4fnOY$*lN7_CDxGVAN2M3H48if`z22&JXC1azLA|81+~wzn3%7qxv(@}y1s^qzD7
zzw^@joO{kysVGsRM2Qk5N|Y#3dVUgPOv6n}CYWX<g%>iVD)<z1;Tc4Dy&rY4``;;s
zwrg+XH@7B)qv7140krU^k?e%sujBW&cl@-duAxbzGj*zTRSs@;_iH$raL!t;!NT1Q
zO?6xoXydJw>^0xn-=COmJ>42zYG$e;Bbk>)msGQC>#5e;Hrs4jTY`tC!{M0lmG#tx
zh~g?_*sm%u-s7+L$+)?+=!Z66)LYxER_i3$T7u^wHEmaa7HWRst^uCe1FH2t_CQr=
z7>L#xSn=WG+hhUY`FmxtSRxnH)Sq!>WVlaaOPNcXx3`;Gva_>eyfu8H+F$=ibAq<W
z*d=6Axa&U>TEd`>u8L{4m2~G8mnA-ND{%i<onxw-%`<QkfZzbAa!dO@-;(<dw37pX
z=4*ZV90D)|0NeD=%nd&-fOc{K(0-X$oJ0VI0-&Vt<8>PrK|2CK00;m9AOHk_01yBI
z@YDfldEw%B?qy4N*Ry%m7WUGPTBZ)*ui^f26?@pinqOJ6`5=3xd^c0;r2X+7eCBgs
z{06_pF(m!DP2gpCi<e~yFVpM1Y}x?+&ozLUQZrkBw2583)6IJO?lN1ilQmxHVDIf|
z@GBA`i)$KBu$J4M%;D@~J={)q++r&a9`-8|KErqLfv=rg%dJj+jFZeA!S#WQyio>}
z1!Y3nP{x?yhTET201yW!=Hw?Swzkf0p*4tW=ITR71rf#mH+$@C*}mU=^1x>|dmKLg
zIL7KjM|@fu@<iVL%7n5Gt0g?E0MN8iyZoIjtl-RFg9I8~Y+}v4_DcC%E?pbskM{@o
zXHx#>__1(I?=m4zDSwm|U1i7sK>yk+ak4OU17k|6ANo(Os1v#!DSy=57q-_t^v8Rt
z+6?RE$HFnaXRnZ_q(3)|Qy~K&weFDQ$-;iWdV`HEdDQPtsoWFL|I!^BAHW|0h|-&c
zS}x_^<#4j`UcitS@(h#*%1i))1OUrD-6Gk_%&qN%`g8)I008va4VSO`Ksd`=#w_Fh
zLB0e)0RW!geSfTp9pnQtU~D~gj!if0^YJ493IXs~Bp(x^^wpBz699z)JkAIJ0SH|H
z2s=#NQ7goRMu{%I2!O%>aPd^%eqD&|ARctQ;}-j1e<J}H2>^?0>)Dkqn~!fdharYQ
z02BaVKHo0se}cCJ@9u~I$P57C<MaP?N%|uqiZLN<0bqipCqQrw7f$%^K3o8JkIw@A
z2FDCr0H}SKD3OXh;_6j=MN0r=1c0eR%vOmfOdvi&2HxTj0EMO`5sk#G6(&v)yGQsK
z5lnv*9|S<byCT%?36v<kNz#V^JmuR`2mk?~E{FgqGzeQngYfG#2#@67P&@q%r3<kz
z1dvRji03O9hZbWG01SaQ@B23ifXo08IzA!MWeNBBoNpg=Ya#)V82}8ED40wU`(Ey{
z36n1>0-!7a4}|0H+s!e}i0MXu_iPdXg#dWEBa8_X%WD1J=T86>1`v!900I!&0C1aO
zWPsQa0Z<44Zmh)(wg^|?VhvZ!;;;n)PzV5SDL?ynr{CC-covNS$P56HQ!J{GiX9<}
zN&plDfaOG3RupK_o%ehK0eFlp4PudAv}loDuZ)ZI9HK>fV}e+uw<MGn>BSFYnd-SA
zE>pDxSf<)6TBfQFw`HmU7i!;;TBv<QVxhJpz(Q@i#6s<(K`+!s?O%T^#1?7~f2)F%
z!gd4pVQf`^ZU@^OV7r5`->Se{B+LP$f5iqF6Na!^$NV8|){*SltfTVrn{{Br5VQu2
zQ}Lm^SqE)bLjVW>0U!VbfB+Bx0zd!=fa05X%L?GTlJa>3U}!Di(~X-J%LX8?Y*(VD
z$k<H)f&-x4s<VB&aZ|D!0Py$?ZAwbYqy?t^f6^mNf_sF?+`JtZl9H0*pq<%lezNDt
z==FL$r0m6lhQ^|3y$~lE;2F_Fe0*xZpvgdVekJ>%tg-@{xwTWt0U#zLBV+z&wFj<8
z6d7C)3ajo3O|Br!P0&DijMV_N$tpMANKa4406Z=yCug)uhNRBS%)}Fc;=WnCHfz<+
znj2Ah6)x%+aL8O+Vdhn`tnG%IYqGOfL96A<myd&XDw$Hmj2Sat=C5z2rKP2>)0G-$
z|5AKOyE)&k+ER2^wWYAn^%_p*I%m!10>>P##kzc5RcdPLGHApT9^@wLGbKSXa^}pL
zvB-ei+}POIrSE^3k&*V%%1;)pSdkTumcq{+8gMP%g(lF(tx=?$wcnHQk0>f3A>mb?
z68QV++-7l040D49)1e8pfktkv6iJ|pL3wOkT-?NI)26)^A0Iz?>eQ*R;b=H_XaFss
ziCY`BY6KE8(YpJ&QDJToZY@SccuJHgQKCeN5~a{c{{!V{pPMtdXLSGo002ovPDHLk
FV1j*zq|N{U

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher_foreground.png b/mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher_foreground.png
new file mode 100644
index 0000000000000000000000000000000000000000..811b17ceed93cebf145b331bbaab3a8f69fb67b1
GIT binary patch
literal 1845
zcmah~dpMM76#qWcj?0>ql5vS8mnSns^Ds!+36bn*Q!_5bw=C@_YDyHTPup%Gv$+hT
zT(;a|T$5`}=1F&r3d5vDW+Ec87>$+LZ)UbXx;^_n=R4<p&-tD6p7XxXb57!RvO7ji
zR}BCF%vKNA9f%ajeibyr6Th>30|4lSt*)e<M|<D&wS_aOSTuC$6k{a;C!o1HcC}d(
zFZtxYUY*fTdJb>iF!g<*tpHx*jbb{UU4ENk93|qGZKz4%dr_q8upXvyd-EwBvsJHt
z(_TJgXflw_k+v>gJ&=F)tKZI4lMW$Zro>W#;KVOTETn{mj1d!>g^&ue1kDgq61aB0
zug;6a$8<=PlWAdtfsQsYxcK-`Qb1QaZ!RiqFtYjWK^tDnNFno&F?guGwt+qRDA#@T
ze$g0hASSEsGTC-uy0gKs$h_)0{KJA4TQYh?yeNt~D8BF$`-XB@Rq<gSe!Ls!KbBx=
z=Ul_sbCcFM<D45A7EIF)<oRR9``f6aZ|^vuik^$N6ZCA>z$xd8h<z#R{v>srNHOFl
z>jXXUSH<@SdB+7@LD`XwnzP$w3UV-pv9z;M1uXy)+p6hB_n(Ydr{EsZ^^~huYtA{U
zpJQ|oc;i5QYg13g#ruxIBJlLlhsxCkiSEJbNsLa(&{Bw_`8Zccb9KqoaSLdprbdh|
zvbJ%7XFZ%q5gt)Fcg=CYSn500DWj)!UBVJ9q$Z~_7%WzhVvq1IsH8s7KP@e3H|IyQ
zq89$<MO~9jCl60Dawi4wE~W7{!;|EJO(!noNhNj_3!KdEerb9$FTaI&QaapFq9>gY
zq;If;w7`p!S@L|r^=6yzIlX?zvdDcmM1mFeGD|Wj+4m-XfYiHxp#|%7&)_{0cL}~<
zc-woBgr?O_`NW@Z*uud0&T@rL!7K!plVvY{dk<Z14AD*T<ZgGJV*AknX{**f10mvy
z##<3%ocHMH)zjT9tpYZPUm(!+5Z!xk2B~*^YEw(wqZ$xZ{)Ci5O=K<)3KjB4f61rn
zZ~?Y~Z_X<x!zrh=?N#ErV`?aCZkh6`6kK$HKCyDy1ixpAH8%-dHJ1||`!M)=iB)tn
z^&2b%M;yzbJt?v5x0OzI0p0-s8Vlub?6}tHqdnzG94YXELA1c`fVSu_3oB@Ue`^x^
z9NoyIo&Wo>)2h~7U~GEj-KXC9&yu!D3Y>O<NIb21TTmUbpgw2f*zM*u{uspA*7$#-
zOgZy#oh}6L(!jq8x8(x0^*qcWt|vfwJx|~QI3?fGvBv<aTm>qsbVjLH2gA(9MdZXJ
zkRp@=)O=}jfIVbdAp;mEFt5r1=wUNE2q_9l^;1R>%U#7=4y%6Tdl_cd%YEk?Jrz;0
zkg5DwW0=`2xBQnEk|pvu_yKD_Rb?!sbLN4kJyNWb7v&a+_8ktg`Poz-f<yJqaDxB}
zKdnGrf1CVpUyuy{>`dt}^B@dTZd-qv+J|%_(3EL;%-Dh^u?(9T$|^sN)aTWzh`rF5
zBv;cuRx6HtB)iGBo*Xt`U$6Rg!qD!DVF4@|($Jx@9gs^2mG8k!dfkN(6x#PLKb(O<
zs`$uQT(K5{6;!H(W+}+xbCuDLT);<8&(l^MZctBT@Y8wfVD}orw=tMRMgz4d+lU+c
za^Y7um&+luw_4q}v7slLLOi0%+gnC}Zd{*+VQ1>L9!wIWt28W{vZ)f)y>4%$^Lo7F
z4sT*DVf5Nkfvo?%7`H_gxXg0L`7<w%XjhXCi5Ej$5XEi+!J)$w>Kl%qso6OHeusZx
z$wj~HnI<Yb!A&@{DC=^C@o`tOQ;`O@J46r&rN*rbiZ~8p(jt4_f%=3rLIrx2UwGhy
zV$Pnf$GRoU9IV23D&sXLt3)Ptl`C~BoAszHk+Cna(}Bw5*5_s2;V{+`3)uUd;AL|j
zBD?m;VhboB=&3N3b#wx6FmcZOiYP>_HxKD#!Z^LhCH=k=g%f9HIoLYK1lf#G41^Ez
rlffVxB&v#8@&U*C>hw9Ak$Ms=<da9XWm|BNaR_X6BfC~^hEDwpSu-^s

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher_round.png b/mediapipe/tasks/examples/android/res/mipmap-xhdpi/ic_launcher_round.png
new file mode 100644
index 0000000000000000000000000000000000000000..2f3c752e5ae2dd482639a89c0a0b50c50c5c4522
GIT binary patch
literal 4658
zcmV-263y+2P)<h;3K|Lk000e1NJLTq003YB003YJ1^@s6;+S_h000sBNkl<Zc-rk<
zX>?RY7OsBXVM|CzlqAF$fuI4yrV-H@apIW~1P4TsgB#$HC5W;aKr$rQECCY8LiVsE
zVd*3U+|L{ZkD#N1AghStAfl)k7Qq2$eu$x_zIydKulLuTbf>etbLw>R`c=KE@2k4^
z*1c8LgtT34m#d3Xs$)qKN$(;l&*|Eb^BiyhXb?%?fO7Pza-=q79V4Y1TT`Q-df-3+
zJ{>(~JaC}XE1w(<UwG)tTT(we8ke^JNIVk$#yxz8XB_3Xct_g^VHbcf?l#zfQB%{7
z6jt{kCDo6SqN-F<Tvfr+`z#$}=?9iBvHQQV&;Rix|9y$2Gvb-|c^>3N87K>7T9MA*
z?S$}_d)aGvepM7HsTxO$t4kPYhZ%rZS)!z<vf)WC>yTVFdk5ZSuOqaI9<K&|!<(H+
zapf2WP9;mV4gmjy6joj{C;ZQosm?Nq`?k;8BcxV-Cwn*Q;HZ8r7Wg7lPX>UIiHNW5
zfUW|#@_HkTfa?|n*?iCIVB~dbaYQI`R=G;6V_9SKnTWZp0v-VU-iUx}Bn0nNy-%)F
zKCh!%R5n{>tp>ix6v~GCGS<LfRlxrt#Dp(EcT9%0sN!{Gbw!=+wWV3~`qI(>^Z&)w
zPq0RvG6U~SHaBfjI)p;RDN<7TWc}!6ZSwokU^Gf**}1MVg2~tF`W|l<Gt|G6Sv*m>
z?yL=Hi`iN-O8%>8m8--wLfGr-UQe4<vK$R(c$@6C)$inep()L>*Egw*Zw!U~yZQfu
zEsIzSb^fInl!~BN6qz=e!7pfrz1~q$x216NW+sLV{y~DGY%!TBs@frgZ=M*@mX%cP
zDlqLND@#KxFle%#pHtq2Rq9g?e$$(kzA{@W$=|e(7_-BgW+W&t%-DdQzn6pGl0=SH
zk+f#hr)0LVi*HF#EC4R^E8lH(;9HZBTfUP_of_a9>qu<_7gcU-C1S^kg;aHoS$&jb
zZP-$89eHz}pfFruO08G8trC0{36*usJf$R~*y!C+plJKT%@3*Ud`;Glj-`%kx6<~d
z)owMA=PTVxJFw(#KIG}>W<J96Rr9*)+_H|?XWm9imgbJ~j%B1MnA~*{>}6-P+W8VA
zD$evaU9{&g-G1T}t^D#EI`5+cG_w3Prwv5Do}1sKsh=F8HD8^g)n9!>7kqq>-n{X3
zsw;K^Jo^sMq_Ag^XDiFQkd=Y5BqzZg)CF~#&+1mi>bQ;76?I14B~3vglkD5dUjGA`
zZ;WnSr3?p@?1a*>7gcFhx~TZrcRr-2E?%NnfBUaRI&}6N{mW}R9p)YN&)xLI`HMFB
zj@4eEL*IPgVLp6!?C0|~`A`<hbd(QuIr!5#n|!D{+MtPqLg)h~iZYA0Gpr;PZjIFY
z^S6w%^x-?N5PFY2@Rjt*nX@#a{8gKL(X0obyzq;|d`B30-6~$Q$%lJK?0~N%l!>yf
z^Pvu?i=%vK1KOfV9{`^<6_hN=n$S3-Kv5`|^%Yf}v>0wKWH5%XqedZJ{LfW7{=JWF
z^3M5izteobUj2<G?b>UTZz{Wo@0{gD+18@(nVlc=x;V?XXzyW-IzeEwdi^OfHTlMd
zD`h#ej^uA%BCH-wetak~$WdpFLdw|xIW?$;H<Wbwv9Dd`D>!_dwkxSu==dA=T;;1^
zWvdMVyn7|9gR6YdUucuI1pt+Quvs!Ci*qwMb{-R;WXvh+$qK$C%egI@te(KNHS~>t
zpLW@>?1Nv>fRZW;;!MX+y3AL4^eYP>QAQ3c%T>O$M^9J~hkV%@@}VtglQuUt!w8*f
z^}5Sserg{vH!yfA`_cPz-Rb>JK>*T3CeAKjaV~<US&1&M5Wq))Nf`wt97DHtfs$cG
zad%eWMR$6CQxVX4T@BrJ`Ui)3_WxK*gV)#ih5!YtRcn7?x}Z15a;iIX1GDVkn5LTX
zQhz1@DWT>ay5rOtdVwj`E9|)|LfP;7?nip>_8m6QdyN1&idM%bXU0{_-O~k%FJrnf
zhyj0GtL5Vl1R#a1-$tK!_g^%9-%+|^|55tX&X4E~8@Ahh)~f_4m}L~5Aj8InY4;N;
zvOlMMv^#dczYw4#J*&IH#RSUB1jzG(oH9yYPJc?9E~uZqNhEL67EeQq{y>11<Xr;9
z>==$ZW6m~)sov%6*r=?bH)ck%0)BK?`L+?@z$DEmIZK{>w!5-K-XQ^4JI{y8>UEEM
zDt`Qh0DSN28Ltzd7DLI4Cj#5C*30CrH!yJVvu0zVr@g<a34p?VVfUwW{pTlW8Jqde
z{9vD5pObG00EWzh%->AOqE%(U*5L>+Z{bNu2p|QUKWmNzB)<M0ch*@y|K+)h^r5$Q
zH4y;<!}R>aBr2+tY~|Y_0sE-9_*t48$5^}E+z9Ae_9~xjIm&aq_JYkw@EHLRLz$(&
zk}0zX0^<X+sCPiXisDDzjd=M30h8a~Lod7VhC&?qoB&{`=z*CF#{*-5we5~VF8{OE
zr>MZc-W&)BD5<7-hg^F@%dEViz9E3e3FKMvg7js;*k<Ra*oO=I)W9ZS8$6l)TE<nk
z@{pjjM!`*h*{Z&{;RN7s+;f!=WvkuFJ|aN4sF@;8%``EL15|ihDPTg{ku?=N8%h9z
z7l_L`YZUx7a2Z&~5n3hz3`;CuS)Ba=!?=?b=H`EN@rj8+B!AP9h7y2;;P^2oAwoFK
z;J0ZN1SnW8%ljfaJ|2D{orNVNWRh;wKhh)E)cM=SAOI^A5F|&ORvjvY&mL}LcrNWl
z(poG55Vgw-zQ1q6gczZSZ3!5cnHS53N^L_-P<z30#A^{UMAQm_6Zq~Og)&<t0cMv(
z@%f?CX7+VNfHb~vU7v=|0URZS0@W5zt0aKU1%}Q0$DKBokV3$dxkdfiaJk|Q0<;G6
zTQmWyO0N!`n=;rD0s3)Sc>~%A@Gb#=dujeXz}hYWV^(Izds74%+C<Q=i7(E%%OL?H
zQeL@@SwW4igp-64&Tj2Wctz3q`<{DlfI|WXKRr2`<ZU|LsAlNdb9T(oA6usxnx6Ol
z?ZZa(0@k~CtHP6FA|?e#M~9I$8yjyTR(qjsBBsoQ@wqtXi@>g3BZ0N`L`;qY6B10a
zH@w#v1T4_virCT0ToGxRa7ADyEE&}+BtU>sUOpE@GE2(aoN<jB5vzB`&0kv0FzyVj
zZ4MO_L4r<XN$w(5EZC>+XaHHJVR0^8&jBUPWd!iJ@O6?~LCK5-Sg_m4Za_xd>_Czi
zKWVf5#zzDoA!y<(4}vZTzE}s~{g?pR0?OV%$@pngfN_nzz`Tgkv?F6DCXfQtFHI4W
zR5nZ;ccQ!zc0~BG*$|+Zpvwy{lfOPZ1Q<ICNkS&!i>@T6?6amK00{wf1k$aaKg>*8
z1m5i}LGfz>@R^!<lwo@lFg7|`qBwlUjP~p!l%(gEHYEYdNPswkPQVfqEH!bo#?#NO
zpgK0UqGV1gqG6!|WBb?eyu9|@5}LViqSanlpA+ED0xSO|KrGB!y^fM+CcOx36OG0W
zt`=r>B!h?CPVy?w_<Cugr^#;#P~`vOoS#Vl{%}KgOmw+hrdKbFf^a5=s(igTw?z>E
z`yWRcaPgD&2E#N$VC&)<I~FEV7cza`1i293El;<Xu*6(nT-6LmfDk{c)>1NV!gP-5
z@EMv2(DMmdNO&w+z22uAG_2QnKDFZ`zR3bJW~`vF_8s{(0gC@8E#qqvbVDDGX=0+r
zy<!O?J8?0TvSfuk3fx<y)ldSkoe0~CE~>T~!FL1)o;@$B{}Eu47h_@aHDzWV$5O^r
z<0XrsU~=z}yGd@D+wp;h5P*$z&enKhMTK=d!S4uAl<$(9^Q7M$_j4?V4{z_zK1m&k
zPYhz8Q6yz~2|oj&(Bu;Wu$9!QE8@y;JgDQh1ju@`j;{rMe(pw&p^T-Q=SIkkg_1${
z|A}O6IOmIv+6Y@>GqtOHId+_Q{zHK1{l@GIBtGE*j$vYA2T%6O>quN2j&_eGb6?5g
z8;>=P0yJ{FhNB8r+BvGA!iz^0l-oI~AVb4Z1>3Y9RiN~K$o@$)ia2&Mh8~`p5+C1=
z2M|I-`;rw!C$!EB_+cG;H5}Hl^JANQo*veb<l$i*I55QJVI6xk9@e1*UQ$+^Bsbo8
zC&x~c?eBCH{JArCB0Zfno#d8NtsA#p9#`{&hsV|YeBSYKHFkE#`0TivLg98^Sxm|3
zr;|Bmy#@z+-uvq>0|^dB7%i5(oc<~oMa8a9KX-Ou6AnkZ-;)EI82Cfqa&cgj9apxN
z1Dg(=ZNPy|O4JJ-^o99vbF5^{8s*e19Z5)l4I4>r?m37o%m0?Y5PM+VII;|K6vu(h
zcjL%195=RA99fp)#*t+U#F1q<?kshW_9M%1oEnZ*gZ_dD)pTT;f`u%4z9To?axcd!
zK0c^1dcSf>B&HvJ&>)74e&Sh@x%QenR*`aOUk4BN_9}<=={$%_cl01%dv_1*Q>>n)
zdDqB;k0f!7WULzf^uz!X5gQJ7os={ulWTep4)SPGNs;>*B~+S_lnZPGMqW8jT1TRz
zJCeXatZV5_W~Wy3$p)s}n0mK50))E(zW%9mUgeko8+j|L*M4b3B4J^{BDB_z%v-XP
zNBg-_ZcQWreAZ@kz!zTL#jxnlF_E$H`Y&wgNlc8a4*~)Pk_G8|S{Df_@Mov)V;CgJ
zazDl^a^Jovf_W9$t1J8L&MPo~>0XsMs%zmgU-t9NO5M-eemBQJ#=_uD(PNJUMJe}S
z=?;>d_MTZW!!oE3=ClOh)raxH)`!@+i?_XA>H7Q^mUMg_QFssh!VrFiZUB>qH6*iC
zkRWX176CqMi@e8+lF756a0l`>%WZGh6yRHu5E<E3jD&vD#ObTa^8DZVK}E&lCkdM)
z5gd4koJ!fhvo!a2>8UBX-1tCS<u*4d_?F_RbLSgKhYlbiUO(co$z<WmZ-riHexpI%
z>(Pb<>8GU!M`Dr{&)bN$Hj&tIPC{sC2<Z}ulcD+;qN462)91ZGRu)k{8sMfa65c|8
z240ST%cFLAAtg!4ZyCCELsY8|Z(ElxUHlDvH3=}o0!diZ&Fu45*vHbSaWAnRbb>1{
zWu+K*Vfd5D%IYG!Rq7S(%ger&9)5hTu<>sdZ3+}^^y{BAl=v#^@bCx{5`teQy*(f}
z_#Wxm7YayPW-Ygr<a0Hz3;%i3=m@in=i7&_n^!QWX4FdKrxl^js5@^1+JZLuV3O6i
zB8iaBBs3~UjD~)O-tohvNwdnyqE$chUWQn+{Ckbp<e`tAX!5p*xh}s;$avgCU09u@
zNi!?-x7|KmDB4>^-8=iM$k8g2I*_2Ct|TM^3OWu$+R&rdP-*<sHDtlE<6K8zCT;%l
zOfG*FF;!<K#T{bI&zt1m4wmO7^&&<&=O1PztAjLtN}i#6>_bA_#PPbK&T_BUH%IVo
zXU>qHfq@YuBBG}tqF(@;NlGKe%wme}b&|UBdkzSK8p_B}5(kc%gM6`*Dsf=9)LGo)
zBBk;V3$T@*4x%wnNzM2U@5H-V9R`aYAIIy4I?A%%(A<HyryBxz;~-OIYYCz$o`EpH
z5P8!B(%;9tBu$yUh0I%ekSxx*zzC58+~}>?W2KA&HL#XM$<M&*!U<K#n^{^%mgJsC
znJ63Yz`O8HZdRfWsEaJ;14SJr(t<CAh)&#|4hxHBwzV=U`T`i12I+hEA1RF(JxiL9
zltrc`ZzXfmc9WF!!z69hH^i9r6Pp;FBkB1UdBWdVfsF6)O!6zc$@9t8(uC);q!EwK
zL|G_PfZmtC3-83cWt+ZJYXIJnj^KL|g4oP5G%6}4Bs@Hp2R7hem%-}KfE&nC0>niZ
zW|PPC>HA=>L4!ud4SC=ZB>atgq20O-!8797fkLMD6=n58*?32gJc|ty@7A|8QR4{_
zl90IoBrqf-B(!_??vb&vu{X2N?(F}bJkTwRtT+K^Z$3GyPrb!=cqSHkkXMuui86(V
zX)oR-wRYgO5Ymos7-7JLvRQU?baYr$RMd?<di00@sOp4!_zutDSpjxOQAQi!4J1S%
oMG_)IFF;i%<-Vi>+wTGYKT1mYx6z$CMgRZ+07*qoM6N<$f`TONdjJ3c

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher.png b/mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher.png
new file mode 100644
index 0000000000000000000000000000000000000000..05585e4c62ae1099cbcd555a43d69d30072efcc7
GIT binary patch
literal 3562
zcmZ8k2{hE-9{<f)BTJT&K{I7(2x0Ko3{7@o#+oSnWlMIdu^W|SkbP%jB1+cm4HdFx
z8)WCjSYjgk&YS+{op;WA_ndpb-+Ryfe!us8zT3HxMi?CqHbFK3066q?HBIO=?$2ap
zrmqXyGqv>day`u(W`1M8vi%>L&7E&`PB^V+fQD6w<ro<g@P%gXnrDjJOvf*Y5c(zW
zef=&JNT~LD?-FIj=MA@tlV*%|6XylP+)g>z>bA4Nn04_q$fz=yPrPG%wPB(~w8tmH
zQGOP+s?vLY_bE<h<8b95c<19Z$ppsD$_NfaL;s(Ac=BRUom2f~AD*CUG|q#-1ohX3
zzwd#tj6$laY;sILefr26!!Ia{;1uMJNmccdn@?8!JUbF>r(Vjg#mYZqLmq5=T}C6=
zj;(c-w!X0@LXyFjx@!IK)Zj$<(x6A=9My*E@<ksDi)XYLQ_tS_r-%~I){!6UFBg@>
z&$s3G7f;hfSwf}XerM5O5Uq-H(qJ`Yi_A%`V;bE2)U|gl+atvsJs1oU78c%UdG&U0
zo!=@U<gPL;)B4)hehy~s!s3(nH=3dzeYqwPTD+25c9~FA_-1~7ewdk$v|`eZP(Ydg
zI_J`pOO-M}t^~Q>>Pt4h)9j+rqbDgN6XJy4*PK+{qtV<9Ei5c<L=~b=$6m^i$_!iF
z_H1syuDDDiX+_-)41CbB^t$okPYPsW;fE7@a$Bot#)oo9Rpo)#5|So%jv?wL`~{az
z|BgzU@i=;7Aq-wEUW{pQLaH9Pdp-G0fg=!u{R!Qv*REdm@^W@|?jVm`62+z25!@B+
z6P#tbEG$sVvkF<NRk{k_t%E581FI!<UTaO<YlSj9+w@y+Z+;E_jvclM@+qH2Y!6;E
zzYK$g9bCu2;qXL_-jttIZxO&DkYAViezpe{JgqY_fX4$d`2a;zQ`2DUW>+0WL&2cj
z82^3A%-3ey2`debI~407O-!CVpq&9WO@aH{Q3@ji`8QKJD+JJ1#oJ+Z#u1=CEGTCo
zNQ40$X9%K#@gY7y9ROtoZUC%cG{6dg{)caG1o?2X=CVB2q(I8wQd1KQL!*)ET&pNe
z4~Ry0=j+j&{J^I>1LVSHD(<h$DFXI&TQezNOn098V{aN4T$5_^SDI!hwue;1tFUMh
zcrS86u&BGedBCpKjiX?Jvt?i=&p_|f=mqTS8k&+67a(6z;xbXKT6w$+m-KYY-S~Ab
zz-(wxeH@dO+cCyK2%*JiHZLYd-<5ptnOAM&|KQHp{oZt*^(p@R`@)ds>f2~v(i@^x
zZ_N(X#SMw|7USPs4ql^Het9S|j2GSvFpzXKA2PaGiaK1VB=~L;@L8Thz(e7CelrBA
zef)s!p*?m;TcEo+8MZq(s`jtV;kRSV8T=yl2rdh9vJTeBSGM_}h+Li?r0HmK04aI~
zGhuq~>8_AgfWdbs8=i^=+R-28LQ&I~QIl-ft?&Kzn0r?tx>acGSD$?MmewUkx9254
zx#Zv``Etn-MFTB4N6Z_j2IZ7@8Fvik-*TJdv~TkP^=Bt#_fbL0QlLWZiDGP|+KWaM
zjv$#5Sb{MY@ldnjja%%RZG<N!=u7d>*#&!1)4n*VAU^4O6F;wF2V!h}oml{BZegeV
z>+;LCq|U`p<FkdiTV%`U=fU-Pxt89<MUszTo^ykcasksL4#I|m`S%yc;~s!R$*xP6
z5dKzVV`L+0;)0w+@9~HHfDKe5GNVVaMzSXnw(qo0<XJuTxWKQBfcN;CAR=(?)9m<f
zL)T4TW;h-Ww_%pLH)e7jKCBB}1^S+nz{yIIrD!?QMrkQ{lGi~5JyG`xn!|ur;_P#u
zSyQbdTX_@b&esKO08k{;i>j-z><m6scy@_v60;hJ8`wAIA4|Z!>HTaQY91Am;2D(h
zMM45_XamZhM->$ETl$-X7VeCg_})RznGI8l`r!cwYV-i!@;|2}E&Am6Vn$_Y1-8#Q
zy^d1E5)$FKG!2FV53T<b<NzZc;0EZMb{UY#O<65mzrCF?V(0IR-7j_Hh@E%Aj{zv)
zAOw9*6^mCWuQUm$8xzMez|R={&a3*F%M#P$#)y7=nj33%5K${bWOk+fQt34`VSoaI
zwi9OXbV|I9_sf-$Y8C*L(`Y!HOY+fD8SAml(b7w&Kk*1^b7Xvb+ToT+4Dr+&(2fCR
zxU59C=pQNYfa1b|eczyZ%hu4O6kk(Nz&%_v!r#_~_^ngW<)RuO2*vwx`eds&3*PA+
z$ZFSQ2L{>jZcI=cP^Xl3@?krCZisrOmYtuD1HY_tPpM)f<8x#+nK)s@X$D0yztDO3
ziZwR3nbfpti~08YlN)8@%qi`2Q@LO}9H(Hpih-5fe)kSjB^KeDxq%tTTytS&SvCp-
z`Z#zcxq*s{cskEb=jg!wDHUDd7N_55rBLeKUC%vDYfE83g)e?LT=aUAM6$WDi^hjd
z^+b#Cb!`^%t>D5_aX_3fLrbZGA68vsyi>8NY9NI?pk0&CfPT!z9V@fpQAnPj7_5GH
z($T5#mmmk8#+UQTefp-q;uayq!Gq#w&PcbbngK~4alYVWYN~rq+=1?-&8L1We!kHQ
z*79_R*8ecI@)y&JYSM2Vi??OP14gUA{4eEch3aqo-l6+{&{>7ERKPspsYHPxdt^(O
z88B!;p4iQcl(;q(owf5u2_b}MqlehqH*Fh_Z&KP@+3CjLM+|P7#Qszi`U5Jh@QZ@L
zNlFs@mv--RHIKejzVON(nzx&IR!1rqJgS0j_^NYx4XLzp0j@g2L6@*j{RcjPmComa
zK~R7Z&wLtM6*SM3*@Wyu!ciRD09_1)-fi#EphP}H6}F2L-@4|d&G@*q&i>0Vmk#uK
zHL<;|5i?2ovKxG{>I%18U#9#MeXr)GPA2?=6(AN-LeDW{AUckHafV2EMVW!|%g#4v
zPw~;M1#Ffr?pzfZ4m1-3DopY7X>y7w{EzELI+=NvL48&Cy?x$T0&AVo5U^2L2wsFE
ziX^F^G)jHsTH|)O*3U9qv#L0fK3uTcJJkOd)A$IG6Eb7nmW0uAL{Gg?B`No6Ah}G!
z&l`?ZY|aL3lttd<NSfbUp;|WRi2P=sxMG~0G#@Ela`jp_BXB-$A!U+sC@gl}=Z5vu
zht3cyo^IZ`WU5_XEkz*1v=Ux}HWm>%yPF+mc-?ZfK!mqX>Z9CoMw&mb_V-@^v?}y!
zgHW=;qhfdQ?im%e);Esd*ka?wW6}enPvnuYS<&CvGX!~Hig4{zfsUcJk<Zk$;vY$7
zR*pP5i!(l23uR#FrHD*&cA8z;shLaFC)p_)9*~#|A6iPKcAp8W^;+8!13*8<vJ$Cl
zzhn{vmDC!PUse=tD&HW-zTP&=B^B@ZWw;0F<xlCiEKiW4I$a3cpA7?QoOLp#ep6-D
zQloZss1lG6jxWp}=Wu%p2Q#AWNixavld68qGx3B6@ZRzztDkROSH6gTJvv==D7sa?
zC?%F^-hV1y!oo~U#MO#BQ@<IEgCBp2;ZI(iY_{aHFoBsX-O5_r@X``~WGWde?77+m
zL{A8bOZ(4PcsyWxS1#EUsP8KxYnNjJY$=u>nxq`q;M^uPN+7tqcSK+wE47I|f<ev5
zZ-1Kmf3o(UM*rUo1{nV|{4acfgADpl^|Y2k!(cG3eQ}3HAg-995>pUxyc2upG;~c2
zbLbcr7B<k>sq*Ls2i^^m<5(nHf4m*8GiqvUYujze(%#nA7XQg@n>I@9;9ux0#k^%*
zdInQs=`-)0Iw1wLH|${k#yWRr%;i;}|MCkA{7AsY?$6IZ)1yS!<!}nzvtQ?0F)E(Y
zUix~$iA~lsRrOIBD|%z07iW*BP$-AD=BSCL__M%h!gnTXrp19O>RJI#&0bckYsLeP
zALJD`@H+cSqb>Tp?@nYBiDLUqn-E1=`uk<f=MOf1RZd8b0Gvzm$A~u3ZJ&u+Zy8##
z7QjO>c%^pqcunPTcZQ;8xPwr>ab|Ogx~MfYy5fPK=EA6V#{$Ct_=y%RmzLj}0C+oh
zO?Gv;M(>h!Q1Q|y2>BG0pNCwkcFLe0uWE3+;Na+NSB5*7*n58_lii{WM<f#MM{kar
zMge&V<N7lGBA2wIYi|Z&Y^vQxNwMlp@5r}2E^fV(_R%q9w~W$^u-xi+AXvG3wA8V2
zmfS|{nDTG8xmI{Z;6WW#uI@Dog$j{TD<H71+pK>jlse0t$hVDpS{5%o<xe%cP`YF!
zeN@UR2n1JFXl(sRO*LB$zPnu3C);L!oLR)e1!=WcGj;6M?htu!#WFa+0=8l>ElyPo
z*u1|IZP<SwvBR#q_-G*;YnZi1jnZ)b7WCD-UKzn1;MXcu!gaJR)<5j0?<b$#U>EC9
zN_JSylD7V%M5pd>E!^)K5{VU1@uBi5E7u&fOs^)|Jz>60V+w4VD$}7Q)6%18xmJqk
zeh3Ms-W*<QuWEC5Q4@#T(bI_F%m#C0jWGQ9s!EHbb$g8#fg%d7YaU))IlDJgYp<m4
zVXxD*{_M)TmFeIlx<a2KT0hh^tUy_F66!*IJY?Q@TiJ3YPHX7s<MhSAA6EirN4jK-
zqE^^_<K|c`!|RXAth@|BOdRk3?|<}SnQ+4J3Nd~sT+8McKp%Qq7|jy(zn}aYF7u35

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher_foreground.png b/mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher_foreground.png
new file mode 100644
index 0000000000000000000000000000000000000000..46b77639c6886d68ee0d5f82fca492cbae45a203
GIT binary patch
literal 5655
zcmb_gcT`jBvL_S)5r`si=*=VQ1v!WiS^yOVL<zkKL}?<R7^);vRHP^b6%Z65^cITr
z4i=;+y(=X`C?S*pAw+_}-65=X|9I=2_wHI}t?a$BzcPFF%x~uV%|sa+>2V(xKFq|#
z#BHFjYs$m~(xqP<hky|{QLmMWiGR{S_k#I@AIlRSo_4b-ZHRv6<WOX~FN?gpl;4o$
zm^{a+3{x)ODYo8cRSPT0(iYjaquE6Nz;gaP&IReLrqPUoJX8Frl;AOq%~2uT@-G1m
zDfBWc&Z*64S`{IR^YT&RbYI7jwF|NbD?8pP^{jQic^J}9N>L(rb_QyxDb1_RkNfgT
zg@G?}{A0JW^MJ5J%#mT>|N9En=K$&IoCsr*mOOTm=_r^}DwG4*sKcszkwcPMT8b46
zdI^a6@0Zo}L|7H;@7Y=NprZ|ccD|o^r>>V@V*8h>9A$VD<*p@{xa-XZ1}W~^RWX_W
z4xSj7!pRlBuCxK!l>1@Eb@O|$V{3Iwp_+Bgt&>e6XN`XG=<C>JN9zjC#86@SSB&0_
zH^|3w>JBv3rFgZUPJk3G@7k#zjto6@&S#;_Q*BrVGiKj;A|2vM8W)PkRrr4=gTP)o
z;1Y+%GZvpS*AvFQPqJo*6m>@sxtzTD!Cs<ZO2DVCwz^sTn}^CdzIXf_Jj*r(D527l
zIckyR!&Z+aEX7<u=WeL8-}FuG%o(+E(>WhKH830Fxf<VUlcJCxCM_v!2WwB2%I?OJ
z=e(ai3_7IC$8KCm^pY-=%^I|aRQr#yOM?`-bqdAB@l1!aImO-jEQCS{K43q4enXE_
z%H=0yzb_sKDSjk4nyI*6?J|csGU;L-=jF;Q9Db<c3=H%T`UmVleG!sGwYQ!v@o$S#
z$-I$TEG8w>>9WP;AT#$WBq-il{xe87vCmKhF(iDovGyw3L7EpF#G2tiiit@y^(r89
z?LV=`aRW~@{aebCG&bFx9pjQCo#BT05c}i=OgwgqDdRbzQAYK8tf^N?_<*(`v#xe%
z9%S%VX7=FNN&JVyl68@jF45etO}(%I>8!dEp?TE~&QE|zq9%*3PF6g#w4~`(&{6Q|
zYyeR4rU0O>o(2GweH;LkDc~pl(`<jUOhk^p4Oj$hk)$sNWB~tw{eLt3E5N>4=@d+0
zq(edY#w_Kv(f9FjDFt!QmtM;yuX~n8o89p>@*0x7Gge<=9?3J6IsRL?*JlG~R}q0#
zg%?&P%%mm0b>2uo5Xb~F)T+--yBU3{jr_3dXe~|{gZ8PMjMFHj?u~sE4(y<5Zus=|
z%+H2f1~%kb4wDYkmCJ}K?iiPa_h4TvrJ9{PucT1=ly|q-1j}MpzP??>_v)qD+;?lO
z9iwz^*!E~vSO(WOw2SodmuZw;ZKtURWEHQQgKa?tSq@*>BV5LjO|z{>pULZw5wf5T
z3VY?F@2kysdnY}b_xfB~)1eRzg^(3u^-Bj9x@}gdEx{hklxq>0-;63|)z1l59j$S>
z(bQL}X^%9nlg+`4>y0v>rp}~|m)DOHR5#W(ZDU1Zg1}Vky0+ZAY1h%0a5%V%cwDKm
z>I#AoBH$&y=r9De(qV0>U^1b)f9-x3iz<`9Ye@|)jiTZN-uv_=VG*aU)ZKP=6s4)k
zil#lh>-W1^-FJL6#C)s>>R*ky-myT&()ic`tB|Ni{Q7HB?mvfTR@<DKeETzXx_Jch
zt-<h!;21<r25z*9eHcy|su)4O(``*J!E5tWojA`Eq+M+b56R?=zJ%TTz7+&J3I_FU
z<Bk*&M+&ST^W9tXSTy8hTD_TeHAluo#wlOC?kCNoVYj}{M=&P(5*#LtDnzE~;;i_b
z`kv^}L=@>b3T_$s6vfYTyk=q5u3bxg*%g-OeSdUR)i;*-v@WxFG7dAMEZd;xf0yD{
z`*Dk0|9pw%L{4jZGxs7$3Ba$8T{`moIoJE^9uP=?x_Dc_<H{>X`~72J9oo)nPxSX3
zVg-^#Bbtgy0wZd{Fkge%g`O)`yg*`Qp7vfm62KO-^2m<s<p4BeDHVjr_i9<1FS?1s
ztV2j&R5z#=x$rTRop|q79?|p+1hon{^$zbeUi^CSgp7BY7Q){DgM$a*i*;3n>e!8m
z!SSgMRrR)<8Gm+m<&`lr_lL28rLelu8$e85LxCym+evzHE2vp+p5TA|%fF(j+raN`
zMyA2-okl_YA6%E0MubSJ8*-v0S|U&~2N;ObL~|H*-MxRO^CO7cetOz@syuG^NQ^2H
zwn2h^_o#BfG<82{ODAEhFkOrCDfAeusi-^h8EZf#)8=a0(mf6}3hV9EK1WaT*%#x#
zs7jZzZM|XvQ+CKiXkU*Q{S~0Kx}a)Ov@E_xRpNDh06&Nx&XP3t7AvSVG2?Ed>XaKN
zm_Eep*ni<+&CgI-y@0*ZzIjiCDr2<l!l05X8XYYn_9PWaHNL>VKb+C5GkC91)7e)6
z0pCj-Us{!}qN~Rdc^Z`Auqs`$SNL*XB868&z3vEoG&IRn(w+3fB7W6sKL5UX6MT}e
z!pz6mk<XlANy3<t*3aFn>zspQdFq4j(f!Wa63>iSb2aj+q<p=0LV(J=)H-DnnD<lX
zVh<NXrnlr-OA?LhuZMXDpCam%3nW?GR%nkcYP`q?sb>Zx0B*4X3-VSKemguWXtQH7
zE48h>n>801G?b=M1T~1Oz65xVzUmIxs+^(CKX6Z4kt8mlFvwivrqz3;&=qc!;cU6a
z6SM*oLvO+juvZUHJ6ZYqEA9K1Mn>JEBkF@~nAlclH&+MY?%1z6b39wqsmE;MI2qp=
zE=d%h;yuG-F#pTkt_LE7ZYQCs?XuN^bhiN~C2yohIrDmMcigckwUEiGK7%dyduj|f
z?Ohl|ogxvTDZO1eMzEfa5xb0iH-#phlWgky;R4@L@0<6MNREejDlqNaVna5P_fWyo
zlRtA>)0$5g12hFN(1u7Pa%l0xhZlW{*TST|ajD1lJ)|4*DCt(rpU3V0(n5E2i7{-q
z%e<y7Vha`0!*7}BK`;k?P>;M(zYP_Nm@M+z>6T<D7ot=6hv=rWDDuS>7#TRe=r7`q
zq){DWF2iN_zswD;_Hex1AeSnCh%tQ~gLT{R>|f{_r(^plIA{Pxjc>O;HQfOU^Rodl
zwAwQlluAV12Tv`dwe)Vg(sKb|C0)#!{6x3Tap&YH-C2>q1avDC!@S3azI;N#c!ZNZ
z0jQ~qv+OJ9_;n+7;)}PV;|~t4fTdC1{iQWMBaPb4rbyY_cCfM>E~m{|@`NVjV7;<T
z?%<)19hXfgUQ`Muq28arHrQcbp_ce1`LXroj!_krKzoi@bo95@f=lQU3j4lL0+#sa
z2If>jCFw+>kGHXV_?0Xq;c>%Qz!itP?uijxJncmSsV7sp_M`bF!{zjY9r*2^(4qPa
zQ|uqw%+d@*XEA*7;n4mo7%u-GcgS#UXDpkCi=koejDycv^2(&ry0+ernU{}QHlIoY
zv=k6LKy!yE42mTbujMZ~EYCG^IhB>Z!cNd_>L~m;kI$0IoZ`S+dh@3KFzLtv9((x9
zw~sYAZZ_H}itNjnN<A}ykHVQ{Pia!k_YsA``6-P2w5RTpb`H|OF6-KxvO}D00@2s2
zYTd5w-hs328#bevR@$k_Y3l@mPF49+G}PA*GRg+=!iy;5mS15aZ{H;T$M+jHi;HDe
zaVw90Up+v>M#ASdXtJ+%fhH@%;22`ds(b%J7z0zRx)G9-)z#*6nToJh6Gc^qc!C6&
zQqk;FPEL(>@(9pos1l+x_+=_)UJokN_Kmyf0PY}@`r6vhzy7$`$JTK_^dOgB4JCYe
z+x~mEd~@1|p$7a>-tDU2D-U8z=SQBL2SQA=S(YKefZb#}YMHHHOXba~kDepf`W0-w
z0RwPJ$0XgGU*9;Jn>%l(UE=WGUypQWTY!VHuj!D<wguN06+zUqm!s_T_%9p&7&yq7
zJk6c488^nYj5{v-c}>i1K@{^!!u&*Vh~U09E%7XUTD`AoVNEr;x+=;6ql?EWXVjnH
zuDJFz)7#to$~t@ZS0F!mqx?U#qkuYup21&9?siQtcWQ1eb*BZr#e7<S8t{D9jCA~D
zgB}A}1P!nimhtv~;GWbiNwg>-jhDj36$%nCgQBGpH%QyRDGi5h68Zl^(!?2IclB`|
z1Vcj%-O5)-U2v88hRfObz0}Y!(I-otw>rQhkkR->-<RCav}#7rZ@J3-g>zp_&?tl&
zFl<o)TEp=)$O_JXE>5+&Fd$-ia7K!^kP;zZeO2G~`bdZgB?19`Fd^8Yuq>Zol86d7
zOdbgt9{MQ25R~;KPcYP?<r=hk($n2(R*K>Hy+^-{)2@C(i0dHMhqi|AZZV7lq{&#Z
zpw_lDR1~kE=t1t~jYRU!4nOBhjm8h<_#f;KeLyXi7E<*6%jFJrBjq3=11o*i9Dh+i
zf3f7XXY6hWhI=gl$#JJ}#cdql{$3}#wLwU)t2Z)}s=GJ&96?wd*k>RBJ19S32Z<ZA
zt(H9}C(2#85wUoHd%4IRUvm~!lVulcl&u}cp%l@j#-qV0YiK+`Yf^($6d)CJqfYnl
z8X{64s7sclF}}UY-sSJ01<~YnXSw9J8lsJ|H#C2E+yf33zv0DbRLOT5QjEgUBKypp
zo~RIa_7Jy?!(HcMU^S^xcWS>((V9C~9>pj9j<unavNF^m(U?oP=3)eS&){8hN+PBH
zG2LP{cI7A|A&hT{*xdk4G|3i)&7nsAC~Q1u-;Z||rBQRF`bQ`!Y2?PTuC6(%<=*%Z
zY3wdPKwrtk%a(kNNuMgk$~JXY{1La<CS(dkMOB`?h-1+h*<@S9y=kc(N7@_pB=r?n
zZ~ly!zu4NUcof|ZSwFL{aCULW`Rd<SV2y?|?5M`}Mc<Yq-E|M=1Oz$crwhONT?2@3
z&yz=IMo1lo>^GuQ`hU%p-pjNSWCwVK2Dq}?g*|Jjb1RJ6Ugb}%|G{T__`}2VuY0Y5
zL*#@^?)h{C;q6DQMZ__oh$gfk?@9l3JAi`={R*mVp*{5Sxo^x4P`Ge5lkGQR+sg*{
z-L#LV_4G`%3jo@0H)ogez|qlmSFPkzJ3yx^$B>#3eDB)cNV-$&?&q&WM66>zr8bzB
zauMg13D^+Zl$!^~Zd9L?U0e+Q{-BdGGLS&KX7j{17N%kQtASpq;BOYeHHX^1Elc=r
zR9YMjs#{9yLlCyNzfR%NMu<UNc*t1)>^Afgz|LlNa(<0KWuU_OhUYD5NX@^L{}tj!
zN&wgnbi%`e^~V+2KP#B^@#&HA+&tgR{-kVP&u{o6LaUfrtU_x(ieKPGL<ZNU?Ikgi
znD|P}0<8sCo$?;Hm%#IzV0vc4X@td<CW5H*Oa}BeC@QuZxDi5pbPSN};}fwXf~$U_
z%!%vlSV+Z6?*OG@U{U^r7Dv{qCQv`(v@d&^)d}N!=lkYSRSh45gT*4y^SPDR*F%+?
zhQ|2@_T-u>n{6E`w0L*yHpp9-ov@lVcDE%`jM|m~eEvVm<iD+^|I6z6Kee{o$wx|`
zJ>SP8#~**ZrL(~j$*$M>{pM*(4P_Xp?Il7-wFO@()8{Qy;JI_3Lka9)?FR45!>+UC
zQ>H|d5f}EEet0ochJ?<brm1xo&{WV1I&U2peTO^p9(i$&N5$?cw4hr-XX9uj^5*Q^
z-9XQVc;Vu1Z1{7&P=YL20bVODq~_xJAzpYT{9S+6?_iEd&N-m#^B!n$)OgvF&)9%k
zSkE1^f6=pi)>S5F(3sxh@tHa$Njs|B_1X#Q*&Z)U^s=pgrIrFx?CZ3vlY@82-*(D5
z|B<NKzr`A<b;!g_S@VoU&rGCJ_B|{)7x_T$yO|L9`V4UhTef8l5mCuL3mX`74zg%E
z!yLK5G(COpT@3rA6HD&ktq^O!Q7th`CdD9~!A^%dIang3?4_CTZNH|@lUo6obT<CD
zydx2doie8N4SAX8he=5yA1I6NTn}TBk3Sp2bUDPQ`JFU84unwVsjwc(bd~E!-|$Kt
zCKMcjg5Nv?wOCl^<~_ge$EL;icrtHDscz#$_?zR~&yGYIJWEAq&ASn!7VZ9wm^@`w
z)W?-ZP}@Vk#Y!+oVnrLDML7J>FQIM{c7?aUeh>{!Unom8E;+H+^~oKzJ#09WJxvM=
zO`6srHp*;-t0NN?8!6J+!3C2~P1goc+(85-CZN-Ik?DWC{=wGZ-!&Tl!A9sp@*ewX
YtL=Ne?>z88!<Wh6vXO4_#hZ`+4d}NAzyJUM

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher_round.png b/mediapipe/tasks/examples/android/res/mipmap-xxhdpi/ic_launcher_round.png
new file mode 100644
index 0000000000000000000000000000000000000000..f8733242448aa43345f5990d1ebef7bedb75aa8d
GIT binary patch
literal 7745
zcmV-H9=_p;P)<h;3K|Lk000e1NJLTq0058x0058(1^@s6=SJeV0015cNkl<Zc-rk<
z2Y6J)_Mf}ENl1l+gakq_(nC{xh)Q@*QJ!K)ibyXH1OWp)sw_1jgd~JiAP_<dL<B*l
z7YoIw_*X#*9Vtc#EeZJW31R+o=FZ)0mM!=0-rekG@BO~{HXrPsojd2Z=lsr@Gv^|t
zT*{?f%B5V&rChFhVT?I4RBWg%4B>s8vU2VT5&(gQ4Ye&mU@{HmDd4J3fH^=ETR;?R
z>AeT9U-$p!+O^O_MYkIrF1pd`Ki95xI8t=IGYt3}{tf;Q-UIKo1QPhp13ah{K)DcV
zI8;!ero-5kR1kvFa=W6`yg?`}UrWYXGWL@3BN@Mvaf^(5WZWmCoD9rHIs3PJ>^;A-
z&+KI~w7g6r8@|IRqS(9?Fu4e5v|1I)$ZLhtHw;Fp`DsMSN65IQ2a1Z3nrHoZR33cC
zVKP$bcf%ZDE((~O>;k1|4`?d=Tjo}OVg+6x%8N&78;&xIVGJncC?)Sc9jSQ__>o)A
zkGzNGhwgp+=X7%DzZnD8JvNtkIyabOrTZZFU<fefa-g*V1*GNu9i`_d6D6H}1Tbk#
z$vTiwAYeO<hdOXofYZ=IouzY4+1Rtv4?wX>cD@W3H#xTvaZNLcLXI0Kh$t~%P@X-(
zA^}tkP;_pU0Ck++1Kf+@k3_lb1e~;u-O0jdlZC%q*)`=q>PDP$<d!q*RN?CQd&#Ev
zrT|Gs;OH-4>c;2DBJZy3qVg-;glmomIxtn-*KU-a_q@q<P%aTXN1q#{=XE0_^m8M?
zdFaBxdDP840wi+JDfv%T;(7`{vJ_eA0VgfLF>lr%;WoHh4}mLH+;=uz3u9}fTt=)8
z{GF0FjVxMe#Ui@2o3h@$Ram4_x+a1JQml~y*AXr-*?i8Vy1FV`UEBAS+?v~Sql^p>
zJ(o3K&uqI$MCHK9EtPJi2<tQ7=o-3RLJ+Q11x<*H`qIGaD4;RLk95|EtQoAGu~bL3
z--6I31M}V{i&0Wxb==xrNk~lUtWgPCk<-uU+E#nMlb#==SRI+HQA*AdvZ7q|=0!r3
zRZ!qbZhmt`Xr8BV9a|zbVS6MgdkacR@~WPpzCiTiR|m1O3dLpF!=)v!{|0H}f~yy~
z7Jdy0YauaM$iVZLqIlUOwZ9mMPND@#&8mkb5^f+luPwQ`A_Y7S0?~@~MQBlKglmIH
zFri5~ZOBbsXS?PKcpL_z>>Fr)QU}*`XFAK1+lbuEZwh#-2t>=*6`|R&F6PdJFfuwd
z5T)h)qJXCgK?JYynT2S=j1bod@r6W|R_jS3gCh#(sq(?E{j$t&QQyAa&Yf(mNTqXR
z(wh|UI1z|aaxjX`*h+MxGSHRNxh?{p+(mkM9z}spF^Dc!WMi}>EspBQ>0Ft>@{@C3
z*Q--jl7!BbYy_PouEl6x!T{%r%mg&^iNc$1Tma@81!{4tZlMq6KIR;xTA%j_B^fep
zjC!SSz^(H>#ZPbEji21K6Nj$-7^`hOMOq_cBYrG*8-A9^?E2A8TsLbo=4_Pd#biG7
zafM8k%`!LOmbv8KiOj#Td$(1l&XpzSfGaypbf&I)KE&r1I5Meua`B$P%uRUM-a~lb
z`OEm?&D;35+a>tQtrA>#?FL?R{1kp_(=Mz@m$sO-R)0ig?!RyuU$_A>OR3CL#W(S)
zpHAXu2%t0)6%<6~`OfaI@fW{dAu@02WS%U(f!F<f8h0b}^dvt5Yv#G<{3ZOW-o1lN
zxc6tRn5U7W*!NLl7I9_qu~c_emLC#Gb7J<Ba`B$vH6PQ(zg}8~@7(?4kx@p*X`=K&
zL~$M>l~a#cgtQ+E@y)V3MrFb{Nx(UH&q2`&4<Yl+Ja!5fmz0^534kBD@33gA2{Km_
zFq^ygv)6A)nx|Pl6t@ba>Curqh_srja$PDip>O1HUGR%ucl@+Th3F6XRld9XphcO!
zD+_Smi8FNZEy*k<itD%YE6iDt>A$J~=lpU8m)^N+QRdZ>Qao_i*B0GRAhA4|$4-j6
z_uq-78?xu1Jg#hU%J)Qv9*%cq6|QTNT&@dDcX4T%s6~Wv?Dt~I4NS=crNOUNip;~8
zi*WOtZ6;*`xBw8h%I=Dm`F};%aKr3R%-%Zy;#!ILdHx_mrYKjIum+<!aWft3%J@ts
zhhj1sl8M~nt&eR5@IV^(63!!b0b$%G7GO959cRW;e0JtnX)*~gM(#gsa-IQ1=C;!p
zB*}y^V&7r2K$m*&w}?Cb*Zw0U0I;=CpS&(fzm3LBZNXfb%2C`$g;H`?%N2D&!~vvH
zs$wz;v4qIVn9Qc@w^COpRc8K)vnJ;mvEg%k;@Wk|GS?HAXNtFkXKkUbPO8jo!WF$u
z+`A-Kmb4BdZG0}(soleRfjGyn?yBqHB}2jxLUo8@q%PtXVhw5%{$NaI2cm2-JjHMv
zAW@acd6KaDq7^{jarPqRW5#55F4!)~%?e{XkqK5-mIPyLMmd@^+W~}{H*sKnw#b?D
z>9J)GF0ukZ+lY&*WdcB*Hg1<J560IQi%iZFLURb&CC`%p-Fu~RYXP~V>sA10_qj{N
zrG6yeHitHT2h9!H&5a4xXIW-Xy@Gh`P5^2`+~T)au1b~}cl5Z)NUZ7sl*j8bHUpy%
zv@hGw3JKOHHTP2<K)Dax4M4zS(tj+JOn9H$y3eFc=Kv7VC2xiH*vAT0=037-mbwan
zV00whN-W`r{L_ZQv)IoJS8)me30!DN8b<HW>?yw$5+bh^IV<J1KCT8J4a+b6dc!=3
z%I@%l9QMh~$#W6_@f+aFvoM++D|199<b4*#hs%}sab*C(fNd1m2m7I@nC7Dlt<H_o
zJ9uN^IT{IClI1i25_GyKxdaUw(VXhMZ(p@sS!WeW%!;yCsfMcoNPq8c)V{-b<xhop
z%c%>vfPg0UyW_amwtb>Ee4PeBDY-yR9-#P4jHb=gQk{>swPuLlAAzAEUvQLh9|b@L
z+Z#19Hqd+&<YlB*3OJ7i5?UhXz)t-!qVpP5_cra33EU?x<7vG{-47K2$v2z|Ksqv=
zPRR817pd-TBGbGHQ=_D;ID1waDgbgC0P$`NN?&IzNT<5D+KmbBtM52JLa=_Z2YFTi
z<b(jCeXI!|(k7fj-ChW?-i-<FYe_~=BuhD^0wC7_5bHSvIia!BUa{Vd@gLHl<ynjE
zk^4~q<jepfR0V2<rp`;I7?27eToHYFqU4-|_KLn-9Y7qd5&DSkkwMO<ZT@Gt?%K^J
zOH5n}Kv1#x;nMF22D~M^F?h-zHC?k~TM&Vz_KLn-89-2;(SO%hcqdWhrJE(RbOh?I
zPh7h}Yn!_i>=1R5!YKhHL}Lq+?x2CAIsz8@N=Yha96icMUomO@ne{1GI8fz6eJBz$
zRa6hz=rhGP@wolpSS)~dN&xXm#d5SX1EcZNMgtbwLcas?$!D5<_CSei;#u_dKmiax
zU}F?A&m|&?)h^#E!Ef&P!mK57QVYbpF$m|UYF7akBE|9`;6Kb<d7%~S_bC9XJO=Fg
z8kb3^bcEvn2J5!CAb@gt9yBZF5amGx6QUj-tm8ie1_TpRcB#@vo0|g&wktlBSf~%~
zY_yp|eWwMGp!0bNMJO_|K45}j!~AuCcVB8WGq$r{dzOa^fbwuV5@r2-ReDi<*6~v&
z4?;RIfOsMuopK)y7}WzXks+ob!%?0<YtZ7Px8yEeb#(y25$H=+6xHuMXS%4~X#vC&
z>4%g^kC{9IFi`+KzC{Aa6UD8X&bpx6y%G46WD%`7f&=9frcNf<13*1C@1ew4va-)#
zGFg@A)Bu9$8_GK-&WZ+1hyoCMlNb45iX#ZL5EetD1F761?z`hA%O?~7$ylWrrOgIR
zhysw757)d=Qr1>S0SH>GtQL7TT>mV4ByoMFWGVoXtJr=$M$;EUvAvI0Oeh~283}Ll
zK}p&F;Z0eNL!M;;hYO*NiI~NMtnbGs&(kB(3V`hSK$jeTXMXmQ9|<P>88$d`)(65g
zP4ni>15k2Kp(6nV#p**uHba17^)`8*nU$viNG1Tm>xWCv`uX`mt6+`R+>xIGzJx<F
zbn4VF0-(!|$N?FiyMq0t!<T=jjSdVK4I(wzLTHh*&0-rC1wb<8f`G=%jV}s)wpSAd
zAdMvec}$N^Xi7}i4Mzh=*x`m2D^RQsy?yo*YXb^ccJw%Y`Lq4BLxvbR3V>t;(8A@n
z8ocy+JHUn*0C{RNGFlM<cN~#KwAWAoBx5ph(Tcmx`VQ<O8bC1_Yul3$?w$f51wec@
zI{DA$0|s}a#>E_fSgP7<(aN<Q34s2m07wB41<>N;`_0}N(H*d%HCNUl0BC;N>P`xP
z6aeX2pua=`h?ys^xrwO}M8JO)04V_CIZ*PwMtui>3-cO1dbEc{0L_StYePO<Ruxg0
z0w8C~1L>l$I}Q5udqOmT#!Q{vf_(UTRRM@&O){ZS9oou4XD}#Sm(m|u0g#Mg{QSh5
z;XPmIU=ct<CE~07`!_(zIhU#eKmZ(Qr3-~yP@oPy38AeVbOwWl;T<<@7uBs@0g#-D
z#AQXnk3ZfLu)(%7%mavw+Gs`g2}f>a2;JJDiT5otUFD&<6EqpO)Tv$pkc?XyvzMGD
zn5YZbuyi^@)wpqE=*kd;mS-JtB!Hk({pw?<M0HjL5W(XP6aYDZ13|Zr8H;`(m<VCm
zFk9Bj5fcS(3Pj6Red4Hc4j5oniY0byuT4Ow0Lb3u9N8F6o&)6^K@1z_0YpfW4|%PH
zV$)I`1t552(*6rqBrPBus~H2`+7$rVv80t9?HT`XXnR+SVZ-dn3~ut|$#9glCYm2V
z(@`a@u)A(e|7AD<YU&uV0w9@6TH&caqb4l{Oh^Gx0E(J7z)_E`gom|R)v5lJ=}z?u
zfMhDFk4eL*->|WO2~hx=Fku2jVg6{)sHcega>$0NJx*mnqXTo#2s3wrLIIFo7JA|G
zKhbk9y#SaPI&^3aivVJAzAtLhx)DmszTk*R%pL%OM^8c?Nvh0*?|)VRq_6Xt(N|Et
zCXfU2?b4--ha~{Pr49`Z^(C*NC_Zx!L!-+bZ3b=x9H~r#Jb0KJ6s;)$(u=@8`0x<T
z0oAMLqkGP(g-aFLx33?04Mq!-5?u9cJPc^Da^|}D<3jMbO><AcaT@>$5g42a8#8$g
zU?C(j($7)|uLDq*K7D*>GI8RZ!Hyu(_5u)e#dz<)Hx`u+UF8p5E~322>^$8HK!iX?
zr(pE<koN%#^}2NN5ep!$ckkX_G?ag?cPEsbd(+-~V)g=%&{h7;9beKe8OGch5bX_x
z=jcQ4EY8zyTo~H}o3pe8b?yEPV1Z?!MRh6TC|L>$3Lt-nqlDF8+M7t*6F`C`s=sav
z9<~24UVHo$-a#U*t*0;GL_(PV{>fg^h5Bw}fyx>GffjcSDHcQ$XdRJ8EnE7N*D$1A
z7USvwGPFq0=l~u7r;^xx)-3>}pFkWwekBd2;GCdH0<GggC?dj_66trx^wI0k`q0$@
zWYav|!UD1U&w|7UsMjm+Pz<zfE1E#-+!#Tf7u^#>s4<FP{gXYdkOCm5vOwjuGg*}O
zB+-2fiUA%AoHPJIkD!L_11S&sFfLvJkQ)F<-+c{xFO%xNe%m051&ILC=RtK)zu~=6
zYL4AbX(#}477N7O*Zlau&_7=8M|E!n|Eb_XP@U-SUz@tICFyo6FjD~JG%k#fz9!9v
zZXC^+?!Bb)AAL9G!Aw{<5julbr`e%rw4(r|CciG&BZCYzBn=_!%qak>+#4E_hy6#y
zP@RXjs!cIqh5M>-V<91a)ReVo+Y!aD`Q08#MSBA%fV>BubGrY+W%@m!FsOcn(pV_e
z@BZ;F3NH7yKuo987Z#(&EuN%04-WRX)_qlWV`0sqH%ucmGdh!6As2Y)t}w#0w&3I+
zPMA~@WT#=*{#^nvOUH()Er5i*uca$78a`no)pcmII<~m4%F!5~m9B?gc;!WunDswb
zc<eU}?~ILfgB13<EbhdYl;StGf8nMUh>_>nv45dwd%i(+9T-^C+UV<1HwHZu>(mJ)
zuT9b1xUF(qA$tHw#U9UHEI~QY*a~_RxfOu)ci#UMwTJ4M*R{8;?#qBo*9S6PAN6_r
zHI$HbU(S%;9smM@EVZlrt+Km#<i5l1W`TI?Gbio=itPOk)p2;Z&E(k-Kx*1}GdLIy
zmo-PTqBqm-oibS=djJTILR+Kk++WVP4}b*g!xzr)q`KvGEEJE)0w94*d!m4VAi6u-
zqgQXVbk!enSs{A>2+q})N_Z%WegjY00d!7&0PT{;0`b;oM)ZHtlh3?Kb<5{|JZ&Y<
zrmT=Js$D;vA~kC68kDq77Ay3(P1bd)|LEj7lcgB#@;^7)s8fu!Fg4GoHtoU}Zb*OB
z0k~41mBx=BMG&ADTLGYLr!SZUr6lVEkN6%kZX?xcP<VZ+W0}aaxy*D!)U@dnXmRpy
zbRsrpr^D86#wR5fz(e@G<hx@g8!f=2Cr@0JUQ{1{^tj1+HYINHyWg%_CDT;th(EC+
z=OvzrgL{XjqaG5vm6$rV%uStl6*X+womrn?IYM{yAw3|JM`|-Pe8No1-XVi&1At(0
zvyPvZ+^PQgt@}*MgrfS4p9&=v)x+7em&_E^L!K!8$3jW-go;OV^Mp)z=CM;&75{U#
zW}W9JV)VwKMO2q|wmy})5Nz>#dp97jt<cQquc#|a$-i&Y;lj<k@!9L*e~D00DCTjJ
z7bQGhXJg?xlLr;`3-vGEx-HsLe0=J>#fnF8jZ&T!0{3ppC5qc|C%Pg(d~V#-Z;1|D
zGhNoQ2lp`@k%feW(VW54kuRg@<p1&$vwn+1=u|I;Oz$~=*<z=9{uvy3G%UCDh$c|<
zR~LW7EpoSsp64Y?h5Ah9V<Ke2HVQPC5W@qZZ?>i4e=%DY-1LR_(Bn_NNp)za$jh7+
z@}S3`YlSr?uPs%hqoQdV+{TSouqZ?K9-^H$3{RDULavsS;!T7&!-L<%ns0Dr{daza
zkNtlAk-kHMOnB}Fm6;}To*5hH*2njR4;hnrt)z@TiVBWbi*GQofi~7MckiDN%OmC9
zP1gI&OTelD!xBMvOostXrz+%tORz$GAE6pGpap&#G<Duy+Q~tNsH{=;C$#ehY(s26
z%NM}M5z^R)_<bqQWY^~z4T(&6igG3t)f1VYu%i0jpYF%LE3K0Wh5B%8Xbq9Moykl%
z`ZIpz^8;4R6R~-YCilLU$=r740wv*nK09E&+cMbi;zm#Un(B<#T@cfyL#)rEgxbfa
zE+t{%jk+Q2vMaQnL<SInHGsV=$S6a&3K0iqqf8*{px6K=)6<qH#z3aOon-pUaqj|n
zlCm*3c}@|k+u&KMGasLNjv&sK2z5=tm8oBTYZzLzqTE(KqzwC<^WtX3l&k*n;yCKM
zYKAtUx^n<=w&cpZX-~y~kmg_(RU^hPr7c1DXRSU|lvm~8c}R7kU4~WthpeEki`P|c
zy$w?iRnb-9%Af$r%L^vf4vn8)fMPRbuNzW^m9A^i3d{{1w~gwC*HN9SbX}%gnLqQX
za1^9Hnmp$K-K2vXP?TA9;33|C8#U=5(MboY8*lG$s;eq=UB;}CilxA75oX_r`dEZI
zqAA+%X{H(yS9}Pm3`c<n)*K?Pe?=W5y68f6!t17%-gX6toaEByLqVu!&8Cp0*3=2_
zhNdkzt$?Tsz_V;M=Enc~oF=5+({#sx*GZ5w@E%Ki$cG+><%wErO|zCg35d??K%^+y
zG2odn{i3F6)92_`ML<XksuNx}8fOGg<?2PApoIvL83oZ4?J0#5Ie_yl%>c`DmVrlz
zxI&mtJe(OkhFcx{aUdT3Z8UZ2bw?9te^2XAcn~QezO11-S<~ng%#DuvNfT1{83vvf
zEaLKW<W_~8Ao8c}LHvUsBd={eJUlw1_onSfi<5;sQMtm4Y;{>QN%#;gSdO`o6TTvA
z{5Y&NTk}w+0}p2-&0-@nPj&-g$lo7ss13Qnj%d`Rk7z9lJ>5ZmISPoZcUkcF+&Ih)
z8MB3~Z6|7Z__Yo*2-jT(Jf=Vt%92>E$&GeU4H~tKo&-qDQh1Todt2);H$D1+>aD>^
zu%?tS^J^UH>fkX3B0oyzU<cUGuPymILj8Q7ktj-giO*Wosz$}7iE~-yz$LkfGq0$6
z^cqXS18Z8NMpH(Z{hTRLC?_ZKW#ME)<QD`WYESbNW1{v!jS4*~q=fqBLv(nZy9{%~
zCw!p^4(ma){(*t*=$i6t>+6b~N313CV7Um`$ZJ4A)CTrvfETIy4UOfdFTP6)(|{Dx
zqHo0-@rM=|QI1=MxheDRs@@zJPq`%QZSg<E25e1zsv~%eB^x0d8c5GzdwD&Ee0|%4
z8CG|E{8cn|%0Z-!$NZ5Y%8lF;NIG{#)&MLsSe?-m534(LeuKI!vSxJcU=8`T^k!?T
zLe>1zuV@Tlgs(C3qh=U@)JgT)z`5L%xtD3UFGOmPH*yCc@%Q*p7g#J#!Q7<Tzo}k+
zYd(vz+OxH4%+@ZTDk=({1w;b5u1(KmdwI1)o}SbiX+lDtQ4Jouo}0d?3@u4x3k>|I
znz}Y!5gxfmI73DP3vSwiGS$G5xtieM9@OfPwIQPwT`PXg1hVW_;4#LFyxA^sn1MCw
zsBYioH8kwKZQP8QKWP@5J_v;7j>;K7aCLseI1afd01oV+&xrn$8#?YYb-T{58?icJ
zY^}VjJKvFNjl5XyBA8ht!i@qWAl!#VRCnz@5Dg!{otv@f4>0R|bBFwgF6rL%-4ul6
z;HG0zG29zB{JkCOC!QL_fD=L2faf~(+1mKCwNkq~aLiewn#{j9q$b<979`0!P#_T!
zsTw?HJs0)i6`IXXSX0p;k{`|qCf-Hy(Iw0itPLblVGDxXi)zrQ9CbveHyLm`(0k_H
zRU@_*H5IEPjT`y0=pu|RqNitT8hJ2_r11~zu6q5Qh1|F)-=dj|d8@P>vfBAPn5bm}
zW#v{4paj4Ok(2<Mq-<(!(9GyRVD73{`!6E*@^pp8@oi7<l?P53TLWKr<~r6|BMmdv
zeyk9xF698!BI#UPM>4voAMZLqHE?(aH(|z4Xjb$+8oddYNgq)3*RKaETRBuE8M#94
zsxn~UePC$-G_zyw!5mfnhh?a{bR7tDXZal7Rq+<5F}v3wcJE$#;kg1H8GyuFB%UL+
zM<hD&;72FeiB>=P^t-BlLsPi1QHQu`^RLmI5}XnNK+%*yzAzYZWrOcl!H-l11pSe!
z_x}_mxfP$KvY1=|695JX|NI1u$Q-z_Qx3wM)Low)Lgx4cn`;L|^ErV+K1Z-P0&r?5
zz>yzFHQ3??vN{(amQBeow+;IR3swmf)LGM@NiS8ef4;98FnlE%HR%BN!G~wjti`2N
zvEcF`>kYsHx5D?p)(4pJ!8cpS^Z+G%7K-G+x)3fyEKDi-V9r@IX7T~}PSx}OoTzEg
z_yuNZI#Fvw=E&yShR&V8hdS(@6pP~^AbGO1eE^G0>ax9{W;!rcSSB5y$oQN3vG%X2
zo_lG6>a_t&RPT;ihenM59F3iN1Wlav3yPY10Zm<S1<hDojAq5$M01whMss3s6JXpV
zoTV83d*Kx{W$t-2dCo6r+_WQT<oln44AtucV?ZXz)@U@{*mree-z``g0W{6nT<fyA
z*JStNsQ|~ZKvJ`aM6gOV^}qysLrowt>FMX!&NDDD!qeAR;7x|0>dY*}6QF#*;NYGC
zjhnp~*rN4I!L8c%fdPNRzrp`8fOG|SsrUH`u#oS7@1oyHzMCdTVJ`ff^g$!IDGzop
z3UI0bNQPEPA567b_*;+t;+uGRdA0EK^J^U#7}%~(ojMWW;o%+0YX@InU%^`GkB$bT
z(*MI};Im9d1jq#0@Er^=4cT{xvbpGk#?aa*z;RMA`LM)WEk;g57*P#p7NHSSHgJ(<
zpnNh~0ub0}!$xZcj+P7{&Ddv}Fc}S)%rN#HLF_yI*&GzmxG0#^OlcmBoO-jIiXY=o
z0<dZ`*Al`AZ72hS{s?9N7sB3KTMsA!Or|gU4lnkd8Uy#FfX2lEr4K3rSe{JbHJI|f
znbj~hg#WL>KI5qe6n(HLfN>L08G=h?05tt!>^+8Z-0u2+wVVqva`YGy00000NkvXX
Hu0mjfWoL}u

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher.png b/mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher.png
new file mode 100644
index 0000000000000000000000000000000000000000..8ccd6c8402537f23642281c83ae8581b9a39e0d2
GIT binary patch
literal 5004
zcmb7IcTkj1lb$7K{E_1-AfTWiAd+*EjFKeh%(5UN0s<>R5J?gxqhuuK3=)(jClN%D
zAUP}uERyf#ySu9U=f1j|s(HKWdAfV1W_qT3-Zwg0D&!<gBme+%)rX3YA=z=&i3p(Y
z<DagC08s0yD$41-__3Wqsz>|1@vGR^F7=lSCaz_v?<WW>O1bjQd>m~wK5G8BKZs~~
z)A|5CHup1iQqd+Sp9*pR3$rlOvK5{l9zH&voYL}bo*<$iA|9R_B)#gnhHSmwM<PQ`
zjWk0x6d{bmmnNI-6dTrAjH{P3r!#)`ZQJuG*&El0@Ni!JpM(bCKkFe)#Ec;<Y|Tk>
z*q|Y}27kt#CZH$*7kD1gL^$*padtjWz&;#zaw4G0X=f8OlpMJyWkEFk3MYh}RQ*eO
zX&KzJ`NQY<){)mTHIzZ3FsV;HS%CqVxz?xx-eO%D>33$o8+oyTW@qxutSt&mH%{Z{
z91LwPQM(dC*6|T%`=q*`Fb~{pV`^5f50X3F4G5@fY!qA2^Pz1GSYv2J5OZXB6Nxd@
ziZN6fqT%=cZhZ`#sf=k8jaL!&W~QDJLinh9uFy-h86M6O7(eF;J2}BD$0MHZQn*oZ
zQ_9IpV9u8u7D$6Je7-!A(Nt>v>skt34ynP}`hC@vyOW5M!_PEN4`v6u%!PcVdLvWu
ziOAtI9B7dv9R0~jH!6d%T5M7(cgxFI@ttE*<hSvfbK*hv?=GF8p(T3p&6jg4UIm^-
zg$*w_{nRq4*Cxn7Id*O5XQNQcIDF{HDxjB+=mdW&uat2*AtmLwitLqKZEO2kPFqpY
zZ4V>gq4OF-`hTjN;#hv}x_7`M`Uk|>T9$7qWa}a@78VwlokbBm)=%H<E3gW1kGvT+
zRA66sv|e=Cs*H>zQ~3sqjg9R9&zk*?c2+F9xn?c4KlcQS+ZWZK?OJ7y;o&o%;qk5c
zO$TAz?a@vmecFG-+gm2yye0G^0u;=3qX~mBqfVRIX%};IbKh@=7ez%yU3Bo_JEODh
zjrbrtI+gIgm@5hUcRLHSOWWnXuhTAVZEYK>!SV_U84X9n1L%~sMv0X&T!0dPA^h;f
zwDMv{BC9;J(|o4FD!AGighxd!2YW<lX`B^p(ydR!sk-O7LW9K>w>YsAe-3|1HSh9|
zPr1NgFupI$wq~s2$CjoKly)vtMCZQ*vfSr`PrL&!>voz|RswdZ{1DqeYRjB)JG{`?
zkpdW)NC9M-!rtP3n);xZaNIwjBiaF6p$+ZlJ7n0@<YYtdT9BO#b3Yxge>IE)3Uy&G
zHjmmZ%h6yB2Jk==6F9bnERMlI`Z<Xo;&Twbn=?$-0nais>(Uc|6F697!_C&E22uc0
z|DXK^5s<wGP&WXO10V=uV1Nh*WbmnEnH`MOJ!ER;ef=e*dn6X+zK#(u_783l4OUFc
zQNWIi?y+{&sl-(yzt)!1Zy(Jzg$2;CXfAnw9pjLDG-0EDYhAq1->sm)5%Xx2uU<y&
zrFIr_nPd*gHiKQIbe~%(>?CC)ja6yS8mo*SW1~n&hMFI?M`$MSwTxlS{JqEDKH8&+
z)s=N9OV(W&bZ%n*v5JXk)l<Z<sOG^;|Cu4E;Cc<uJ}mHDpQsgVdQq)bW1(<Nn@AIQ
z$aGf?9c!sW5;P?79zMgUq^>J;`%#Ud&Lbs5rPDJhZgsbj4PA%zuQ=guX5=`aqD^IK
zPV`QDEDv%vf95a{scTp3Mf2YCmGah<!=bl(%;jK;Bu{A>qW(O!$yfZz6GiRUtg*fx
zE<YUZ$*5@+sCPzg*9LArm$FMp8g;?3b<ynOm{0rYDQafSS35rEO+z%b!kPOn&Eggq
zApx14XR3zkGp&WCn$HR=-Un3N?Oepp_ViNuwvFU^81v{xQp9&s7)w|E-OgX;t}ZOz
zsZ-jfH~co%eqE!AColD`asd_fZk=9Fskkw_GBLftA`B$rb~p{gNB2#hT16E%pL_qI
zcZ%d}y;VS-&h*4NASc@V_B&;Ab>;F)(FE(Kzn#?$KRK_3N#xu;k@1z!)lV-r8oxM9
zH90tWbrWPi2Xm6|U2A=weEV4M|3Grc^)x30&7-(si#R14AbjIwKeI~g>%8anX~qG8
z9^wwx4rfFu<c_MVo=R%ZqeX$vZJ@<ozD!&m_qPAfpfpm@a5d26`C{ZEmOA8>ug9z&
z^=ikl<S=VOjfJtU%Y($F<LJC1eUqY*gS8hR+W`DsTuqwQa~(-%rsYajVXVMSj-aU@
z=GAGTaoJ!oD9F<PDedwlV4UE*l!2{O^Otdi1OVBO!IWuo=>WncLYJGH8tneQudd;k
zSZBfaxWHi`*QiC(g%o%$YQJ~A?E3vI?Z*+d6W!HBTkHqp<?`s0{gib@Dw;@vE2EWL
z1kGY&co)|&+gdsTw0Ivhpt1C<{S`(2ZL2T<WD8)sKE`#6*v*jq9SJC9qMpO%S<KNq
zbf0u`1L3b0prhqT4+5UCU#PwLi2ZK&TP;Q#!m^FPllDl}Y-Zv566drmN~X5iGY(dW
zD1-y^H^#B)e5K7MO)cVs8atGbQ+sevx&)gQ7Ne^DelkULj+*oPAYUmC<VOrdGBGj8
zk?IQO`h2e4FkXKbrH&J^8`OUw0Sm8<@x+C6Zi7h5-o4*Rx*XEd7kVd=($F}mxY-=|
zko@1;L<?D;$d?8(((A!*`IF>es0`2{(q!Mm8b3!pZ!G12)O(0od8(RJbQ&F7orvEC
z`5D0jKfBslSovvxI+wQ^d$gFMHuvVwAx2UZR2&Be-qhz&vQ}+g?dM%rCEcWemM_~E
zWX4EZdyMkrH2hr3Mzuo9j?Cbw)GxP2S@kqfy<bp53@SThzkywe)@2eBWF?=T-l(il
z&egn&3FgbcCeqI>QpM=9M?*li8K_$|uRv#;#<7AaXg|ezXqicDJkYWQd;~Y{w2yEF
zUwC(_0?Tqh_6x4+vz!>g&xdriD!$MDFL$W%<R{F|-fq<EVMU7#5LCznCYx1abZdh)
z`h?jpMGTFgNrP{=9ie$8rSk*&^jJF62SyC^5U2?{7Qf;M-Yr=iSdlS*nW0Rel!Ev#
z!bcK=?wO%VDGnXxRQ||k`tHF*b*>p?bRUL#6pY*HWNw{WBScCcy)04EB6Jgp+sC}*
zaeFVo!1Wl~KU6-rSNy%+YK&moRBn5{)nja_ONwTTk_@7_U~9Loru6R`st?QuJr2h0
zTCbf_G>*^YK*$dsdKsh-d}JKm8}n3vY$apCZrP3yW!}hM+~!z(CL-mXx7h;uditF^
zd{C62xWK%#bH5Tl5*^Uo!pwnu=s<jxG@xLeQ(NP(atvr49edl?Eb!;`GYEEsqqyZM
zAExyhwe~mt&0vH5_bDLr-Cj4f<5UWnViAY*YDj>gzJyn)#$HcU`%_J|4-htT6UdN6
zCuJTQ`g-_WH|N$11psTz@M$ur^NDGeh103zbimL++0NDIb%N<{F>Y#2#>EZi)*i0Z
z-8ZjJkS0z?VzQ3|UEX%V+hcTWhx^q?#K6UP;?ZsW$5nj?`pVJhPLnGKaFBV^^RQ8z
zcV?)(Fn6!ndLR>8Zz!%%%ZybPKazS=(fo#pNh5cP|I}WO_MCz_nXs#83a3&BcG`~q
zYYz#5i;k@5`Q5u|B#OS{;+#C6T9rQgB<I?U8F^e8Iud}TzH3PJZjlDj!4zI!zkqJv
z?r49yqmwI=4i!xY$$#4luYoqo-Hi*p6J6_6J*&mi1{x+>7EN&^WCgbb@VUyE2bp3j
zdBU}0n8{kluzHJ#M!lP}H#Y58h;<;H`-B|{;T&VtvR?I?^=S7qvGs%lL2`(o!~<sr
z$->&ff97m4q&CKaR1lc$3>vss$+0DfQ+cJt>F;b%fayp9ex;tYpT6G`9@uAgF<SxI
z84RAE!|gzB(VUt{NgNKyaKv3doPAlJO|94^kKPFaDwdjL^GR(&GcmT{u`6!Jbs*W0
zxS%$m_j>7{c||mN&-Q-Ynyw#3s$^KH1F5UvT1f*<{C4~Gz_AVb6BFn2b<?X1aPpE_
znHzC$ZK1!1%O#lUrM#ob4nAbMCgAFqi}w6`Z#}^}_$i6V=z!Yz@9?gOz9xp5)uS{R
zN)0yh=JMtoiK>&=+HvR0mAPO3`<7MK0LPP51IPa5bfyYB=eDPn<s3`5=^JnJIQ=o{
z<k;&ja!&MMB=i(_R}$s2`^3-V*qZ?V{V$=CuzOE_)G|km@p^`O7)cUN$7Pfg&3i|~
z>(`Ds2xfmgLe2b&_0byj8v^rwH2nHQ>KgG}=}G3Md}f0CI#oTt_Bz{g^i1u1_|0?;
zwfnWwD_`&n2E;}uYm^GCW}fwrwK*o`M?9i0PxX(?$%}idwmzEi$ZS914ye#_3;0kY
zdRSxY>Pequ7<KVJmTUQF(zSbU@&iqi?dOm*^^cy$U-GN%z55G};8l*f3*>-<gl}9&
z=!o9f#a1T)H6JnkE?eutWo|{VL83Pdn6;?iET1}Ck)`4^Z=GCm>~Lpk&{G`GS)wK1
zB!d;)ca@$q%s%W+6j@uM&6!qtoL8fYrYo1E`fHz6(4YA2a?E7EFph)kTNZq{#<$rp
zM-r!Ny*l1#ER(R*d+awc|LR#f)_u5qv(tqbxK435U?!#F(9^d9*z2)qaqTP?2~FK2
z*2-eT2g2O*wda)a<4&FiaD-TzHt$%~nxVj(#|w8%qPm=@hg~jsFK+tZ9ca+6xwJbm
z@|ECI+UPknkaV6k@9QPFoL#W_ILqS1nA9G&$s9#ZJ;#SWhU(vcm8{qBzzrw}|9^eg
zA9n9Lkj)^I4HU-RExLRTfEN#dSI_D5gER$&tZ_fg2?fOCqgI<SpD0ns`>V}=<U*B{
zrN@T@G0^?vbq7db{G6r@y)Dw~EeZS(L+N6(`nVli)756(1v9C_F3RmJUNw3jX@M*&
zaYz1*v2#1SW8P{9nI6!x_l3C(D$_CR>+9Wsh<tajPXaNo2Ict;hW-6jsj$OWzbg;k
zN-k7>*s6WUp(0*=qz6OAT8-tZwlbhvJ}J$p;BB9-R^ZFA0#rb)$e?L8UxXBw-(WTm
z$k)pFkw(eK!{a=;aZqg=yRn6F{h}1&n(&(~CedPikv41CCPp(;vN0wV%u5d3a~!#$
z_3DI_e7MbBU<o}dpn=TMxQBOp`<ywW?T38e&n<7Q7n}cJr%*6l-(Ijbfg>xz_Px^P
zD+Ns!<?+c$TRpc>nq<Qd!4;fX@*-`KJuyV^b^j*UX~M12(Hs>@>vD9aE-m0|^*efw
znW@JpEUv2L*jU&3dx*aKSs(I1{hH-XRLS6Jt3oPw$^Hj^5hR73c4hsP&6cn*%rc#D
zS52q<;{4Pvd^jvrSw&GviGTLGivM};NnF%p>95>`4B2%Wfxqa}_eD0^i9f#SJmVEP
z?ygLp(mg*pT%yf3uljxfJ+~g2lWg;^;F5JvhJM+ZY80~d(z`=GQxo+pEog_|iP0<2
zpT5dvyM?=)fk{d0a!G}sXutf#cL~1m!DEYaTKC*8My%Kl=ZxPe_F6dsIUUDmo5g((
z6_u|Z(3_L=<@3<Yz)=ZQx*>u+tkwi=wG@5&dtXziH=qZ8t+lmFJaCK@`J&{Bdf<~w
z&5KT@x6<~QiZEM#@}$<hrh|ggN7-2;l)l(4vA)4U1X))22=&`!zH{5r@4vdwU5_{3
z&e8`M6q&RXf3`v=Yf`8i_!CZlpeuKyf;&#|KQR#do6~Oo(&o6hEAZm)5@#o`c9p}F
z#1<PiZtJ!__M%h;)~4`1{=19M<j&UC*;L0X+!Wi_^Izp<1Q^I<G6cOgHf6Gov3@a^
ze?&aP+79|mk5&tYx6jYiV>2A*Nyk?yvMx5RbDf>kV>D9*m%Xrfe!)_m=2pWM-ct%y
z$-aGakL@s*mpjD0U3&Z+c=8+>>FJV$WDNJr!`T>+A+R7@{)cmxm4;J7cL#o{V-w^C
zS442Kk@zre(-(G*qWpHmQB>Tb`Py01OG<7>@oWmOHLj%+P3Lh{5QflyspOtwu2Om3
zpSk|4@65LrcZ3NCUiG((wS1N$;kqpo*L``ebZx9aZ`ontR_H%fBCq?*J(63ON;voH
zUuUPa*l+Tzy6*^ndl{r)+ZQJP{4V7*DXtx|q2Jsiv2|S2Wd(`6?;aN>D3Y>e{-??y
zi|%RImi*mvALKNtlr6HM*IdAH`fn+lqZm8G@4P$QJ}*3zw3(`87|Yw^w}=%!!7cVz
zkd%1-q&IF9=5fPCmrQW*ZGXxCyWdt*@JmAEO8Ak&O|m}d&H}1RT8d@zmO=jkIlSyw

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher_foreground.png b/mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher_foreground.png
new file mode 100644
index 0000000000000000000000000000000000000000..06b982e31241cc23c950a6a5eff483f05689ff9a
GIT binary patch
literal 8278
zcmdT~2~bm6mk!d3vUZ~aih=$`T#!vrKv~i@+B8Oq0$~?WSp?YwvLp~(01<%}WtY&X
zfCgk$AVLxZB%(l2b_m3f24u@3kg#W7z~|rdSN&B}Q#DgFRKa_7?tSN;d+yopcZff4
zVJ0QMTO0y`NSXhB$`S$*DHHx~-2zbFS!k<4AU{LRPo1<58<-hF`wY)VYH?!W@I1x2
zpgz~l;yaP(g9G-4FLrun514mfcw5!be4*E3xBJX(>(_T*JsEh~iG3Dv_2i?HPZL=Z
z=|2ROLkG)qn&MH_Oxm77!jFcbVw7^IVG@)sds2cvi&2K||IJhuMyAiAM;7MOh6O)w
z$~B6LqNii|{5rE!|EoU*h?{rroUQGUaNWBXBARjM4nz`o+oT;G|F^fT^^pA=wDq^Y
zP5eLdN!!zVBwj>ci1)rOCBvU^-fYR_W3OVxZ2YBNo?TRCIab*qwHljeX7gglI9HkN
zoL1Iq-s_?R)6m`9u9OPIT_Lp<x&b_=gN8JyQh%8-2Jhp(@L2-M!1uc|jLA-q@ttLX
z%c~s;NCvKAi4)*gTVfPpG<Oass!(5YuZ}yGeF;T(UFa!D3mE+HiZO6}R1edA_+{WS
z$+$WPRuPE4%7UfR*F?6Id1Y!4v^Xv$-r~K~qO!oE_Oj*~6Qv-hkUAMxNVds@Qm&-S
z#Y)w>&=R5s674?9n3}2@>)jkJGF{PjjU1)YI$@9kt9EZhc310{NT3XpZ@ajhjN3U8
z5(DYc57)!^&e+JiKbzdFwX`rfpuuQrvX2%io{J~LU*+zW&b5C%umz7#@1i1Kw9tQ>
z-Cqx}Ia8Olw52geV#?<s{zQIr(N5+PS4dZ@V@0Rq<@#!(&U!X|Id)DYp;o}vKd6|8
z-|uB~b34Kmu#?PvxadSmnJ)&P^&8BJa02@e1>noBCb%mTc}njD@?0fXZ7l4tG(ddW
zv~1v{R*HXH(_}JiCaxX5b7q^OpQTOr<>6_g99-YYQm-HEFnjJLo<%0~FDRUd7E$kR
z5(|n*sl&_Fa(CgmCqF|(?bPB+Pu)SJRs}xymP_<6&yX&?oY#94qE70i9KK+Etf+Z@
zKgsMzR|x#kLeL4tRJ=$3f~*t3>cU7{46Y^*Z?8{!6Zv`iIG_`bwUV(?Ngy{C??DUu
zXxGp16B3Q;-Y`s&8t0T-Uci9$PVPdl+5$%&p_qbCdz!1}xa&6;G!!_f2%dD}Z5i^H
z7Dc?n>jS9i_b*H#qJ0mohPs)AIK0EBag`n=LlN!hbH>!^VSg3N?iAj0o^-T`&t&+^
zz(T3tk?;XZmIGi9&&Ts(Jrw`Ir@fRZ9+nh-fWtn+=leOuLAI0~8Yj75^Zf<j*Br9d
z6>v5DCP`O_6>@V1<ScUMcfvpEfcgvY5WpzEJ*<%;TePE}fBP(2MA8+%KKzZYRu2&l
zuRoB;^<m&-gtLSXaEXoK&Ff76P8Z7ZUt9XHWoeF;lHbKLb<FeYkqdsE?VPT5+Xlf9
z5w=oH)LohCxyX34R1m;_Jj$rXIY!vGt&#|%jH-0+#?`}mF>GHyDMa<m2kNR1+AQob
zL4!}#k2RF<iJ2NlI@zEo;Ui95)askGs+GDxmVp2^x~AlUAN<h|OVD%D=YP#w7A53j
zSCm%HnWEOvKarIKSqqVi7FXa4zqY;Rpe4M&(hm&o1XNM>m=_xkZ*(c9?NF0ek{D`y
za_Q?dA#~<lliO9JVR@y&E?JW_M@(tfMVkr&Y~fer!cx}QdGu4mxtv|KpJ|27`+sdL
z$8i!v98tgb-nLiGI=+VW$GJe2hH6xbS(7c)kUYY&8O5}X!QymGwlP$+Y|^JIlTg^X
z{45TY;IdWkI)98TCZAR7>2~xX^ozKQGwH|*50zS_D?F45-sBz=m)6y)J@TKpEtT?4
zf3)|C7Fxuy^;nYn#x9wQxRG1)iFt%V?0XdAuYqBhf>Qg*>!!_Ux4`77`yqqBI?`RE
zlAgCh1vtkbEL|)J%P!{o__eNLm`az%KTjOOnvKx7%R9-Qw&|R7sBlT!^UO<s9tfS5
zI)EZG5Fr<qs1jX-^VqqtnnE9UJ@iGYZ48!<`dKm=u|Er~UR|cyk3C9D8!UQvnPuo)
zB9#y%AnR^#8B=O&Tm3+z3!3jmdDT=Go2XM)F%<EI&EqrB_cLB5ge!<Be>zVLu({&T
zQD0*4oCZ>3ynKGOJ_|{+B^EqGQD#vgX0NTXkLxK(ul_Z%^K_2k8ed<TjR;{4G!89N
zpFB(OVT}`t1lHcV6!zfkvh*k!yY2@z9g8J3cg~*0!9z|a&L|i`*Blq|k6WQKRwb=7
z+nySueBkU~)>bTzjFD91&aSZKudr&y9n~dAaJFV5agtgQTlM+G1hP@&XF~ouL%mlh
zXRdbp)w^hC_H@$l0|xmiru@w@)cdbrk`TDV1GH;NIlr&cELYgag6DVmj#X>l$wg$L
zD7o~i?FSO{LTTHVKPC}%mmUHB1bDa}?sw(a?~wjpbNy9G*7Aj9Aoz0$PvC?a73Z6|
zW8<s0#;Ojv6d&nsM&00^X+4T<SzIh>e@ul^iOp=4b^pEil508EeVV2ihE~z@f}U9!
zTKw>Jytg8wmfvdU9Vg3nunYa@DErB~4*yj}s{nP7o2;hT(6Ml8KZ?R;o+UmL1^l>g
z$8I)dm)1yLW2l~UV0yFhWp32-_!d-V+iGV|<a5k0%&$MYzm+>1&mUFIIlfjW7rdv%
z2XLf&eC72((6^cTwki8er9;N>3F`ehHNwJRc-<N84}@f#k~zl3ab-3xHdZ&saCm$A
zoVf~j3;X_FWqhP3vu#$F8xIw#R<oR2Zo+NZYOi^=97~RJd`y7V28Sr@QKSZs)kIT>
z-TP1!Gl!fOg_~g(6EbQk(5lrqsX<*qQd48`TM`r^+{IBpjNZi_39IRwNsaMiYqhT&
z;X0vpmbxu87_F*y9}26>(2}GQgPsAe&;5+o^_N5DiWKGBWsIyM%*Ub%d|k}PN{qtP
zKhj3UYR~9|ViW>W2-Wi3Pm2=AObm7dt0>gB4=z09SgN1pmbs@8<k>KX6arIM>_ID(
zYL<GLwGghFS-T@H^9qx{9h{C&5KcEbpF&74T<LoTpTRvSlx>3wFo(vHa&1!xl_vgt
z8*o14`YuRM*2l<eY@hBtTVV(Xd=YXO!&<uMr-1eJWE*eYAc@*BP=d?U8!4@;VOEFN
zcLU6fp+a5KGDoct!x4zS-0l)#J`nOK<`7_{XPN@1x9Xlz?yd&~lXB~ffpOVadd&V@
z3&yUbbt4C6kGkxXbqPs%TStjAYxGpJ5lSFrl(wgA!Hs`c68G*W`rv%<Ajx`|{7pbn
zJCZV*pjUSSQ}C>kHLds#cG@H{929>!IVzu2eihM`TpC3vjX39Bxo$TAOP+F0>NU)n
zmHMl%?}Q}Z+aEO}0d7U!SXqwW^y$Q-Fiuy%!h^k-<nQd-jLZAs4Mg|(ys?1LHQ(+S
z(r82!g%Pj}e|35h2UI8P;_^x&1$B2hy+v6KlO;HKth!0}6HUJTiWjREJ@e@6<2F$C
zzSD|YXhE=F^7L|>@1>R-PZkB9MhH1rB}w;69ucd}r`9%9>%Y2PE7vnBV=9N<52~)f
z9TW-F^S7lz&fm2EKTmD5t2bEzvJ!Lh6@y#QvqHT(@#e}R8MSz$gj30D=e&>_Y#cAz
z-EfX-3##76q26OOLYQ~4Y>ev5d6`%SSK`T$qb5I9<rwTDRTGZU2xXdrad3+tc9bci
z(3XHR8SdG&d)s!aM~Bvyb8Zzj!t|<cPmS4tYEP4L<*VE8(hKItmbX;Kdni&@7Z(UL
zBl3p-0FiHG6VEJFDo>@F|8VTyAmUN0^4%XmJ3ScvcGe~?E;d1DoCieMw9OyZ?HAA>
zT)s?k40)Tk_U3W>ZrBrkn}*o8kJO>{Bi%P)zC*Ti_GSmDvFxK&--VHOCA&B)FPTwv
zGeZTRCjtiH$3Qbz6`Q7C5Ao3RuWMR!sj3qFc1mfIW%Nw4dT@?XOYj(!D7v`Z3W{*J
zMh2+fpKSNk%~|Z~JGJ2CE(%&}40ZLdB%*5O@?f^Z_vskcq-klI8!MMR`F(>N0<$p`
zek2rgcOwJ=&b3*8ZPBt73e#G*tJ=p)K&@jgS8(HFAZTdf?Cp*N1*$)6h+ei!y6@Kf
zIaT|Nv^@qJRU1&SRvdUtS-xH2oI#?~K>M>}5Tl$=+YNO;n0UOyg|;wJ5dpGHng5t*
zSg^8td0(QdvvDnGqGfK0rGNTQDYp8srw-32#(Qo7l`M$DoAvmAaJ!E`PJ<4^N;hqI
z*_N`o_85%;?8|2fAB!428;IMekzl~!DH55Z10BJ2<g5Cc6w|m$4=oXp-D*@2_io20
z!#^hX9yQQYQgW_7=9J25kOJk?>~oY;myn`FSQ2!ODCuE-W`*Y)ZDc)w{rC)e{9&B3
zkC(7i+pJ;K$b67D;%d8A=5}NZ^u_6$#A}x>E^yn4jE)Y@Y58)Xa>@hp;<Gl5G?!xg
z^vQpif{Tey0*(>8Z0jQqqmc|Bj`NDM&`Yxkf!FA+&)AdO$bG1vvA8K_!lX{77`SaO
z;CcmWZ6hRGfrK63=6s9P3iaO!!crPMob^|ba29m4?@YjX-wEHDfXjHyE!UIkFjvBY
zD7f?C9J{`+N&FkDxq!QePF2Ms)E%V3R57vqG2OcX;lPi1QfSujPL*_@F#U_|g7OFb
z+t<dzRCG6lQfE3F78+svQP&Lai$4D&#xo3~quz3o$K5+~AKdEoVPeTKipA~t9ddu#
z#!+v>T*?o%Zu+*X;tXMhmX{y;bNA`10X{|;)M`~g_p}Zu@F%t}!ML=U8F||N&X(UO
zjV`6$;*}oX9q+P#hV4&sKuYa;dr-+@r4()h#;P<6=Hm0XUSFvz4?Gh{KX>nmI_9Kb
z{oMlpB8Hqica^Q&>UQ#O`T!;21gJYW&HpbRX}^xsL!iEnE!ytf0bI<4TghfbUxbzJ
zJ-TPACHTT!yZy%U-%bS>7y`oHqutr>FqiJ?n}fpH1h8Go%lAoG{XHw!A&yq3x?LK`
z{&CaB7W*W{ru%tGDL}q&_5gP+?&tLLnR4y7n|}gLXP1~A5SUY1i8D>+-)H461JMWw
z7L&K-bH9Zz4+B^FaF=BY0ozp}4}RP*zUJg^I>s+ew0I;1#G~)Z=MelcB%SHs7m4oj
z*Lo-Ja|~zG!DCDF;bHp+a1*+Fu%segZYvlAM_uX(SWfW%SzdBd3?lsn7^Zx5QWwdc
zeX_slh9&nS9QmX4Y|w1&)Mm;4e&Np!e?8a6z39bF>}Cs{dOr|?9HQS2SGd~u4Wla#
zqLg;;mJZb@hWpxpTU@hZJHIsYf!)`<3<dIrXL^Z6hy^$p^+caq+OE~b+(wO099>(9
zWeo=)#N~a#NPgX%=;wEOfzq7TkvWLYnpEX+=`fw6vTMCnLqLeWdqf%ZEcSrCKya>;
z?(f$I!-QU4X+TUX0N9G}>MaJv&}Cb$U7b;Ckre1}@UHA2v}JCo_g4BqnX&LrD%i1H
zlc8lCrY;D@Vo=xCb84fOwY!cCxI$+Is6SA@pmP4zy}{jq1_{Qd{~S2}`&5s8uC{HO
z5`dm1085cgn`Vp8tlr9Fx7w`<jI~eTC~hrlPZ&U1!{~IMXH3$_GGIt(fz)2N2w&)=
zQ7bg7uO=gK#HZoSL4skRRN~kICDZcwKQp?9Yc<hY5h(lHZ&|-{;!?Z!u%!88ymuaL
z-?M{g{o2zAUxP4jgVaT5sL6JcB|QTsxeZ+HBV>6tHM@a(|D(0%z3~<jA<8jhGdOqz
zQd{bM&LAS<=|PlqQevD~5c)Ex;C%7E7ZTnL?W7w()|WXj%ni~Io&sQTpG4H1d27K_
zB<=fe>UN6NS{_#hPr+`XF~7(Z@qU~*`~t`fjD%xT-~qs;?P$yIgT`Lp-U?cloS%J}
zdYz8;!--a+XLAEZz8gGMn*|!yrdIQwTv%mLULs;XENtkzF^lIe1BDh)eJ2=;fy{w2
z-GyarmqM7Tjrbc4zg`6NkWPZuAf9We`7S@haMKSPDyWrQ0zi>7720o?tAVHxCin>a
z%`b#~q_q)XE_Vsc@l-cfF{>Kw30@Q02ibZv>Mac~SDRW+p<si<7g7ixWUi}bp@F&&
zs6sNUBIC#WCZGpfXml^1^`u)i%ZG<d!7H?l`S0^ip(uq->6cl`e)OEHEZ~}dsD);(
zCCH@*)&SR|f5a57hoIVD(Nj0w^Uwy%ji!RKOX$Z0!Jn6r2FFj}7Q<_SY<cD=KRw-F
zQe{^z;x3{~F#V|;xOgqgt~C58<G40_bNBjw4Y5K5NKMD2c&9waal7VO0R4Qog87t}
z#|Y@0O^qK*!jkMtqr$88l-Lsbt&Uj>EeEdbq8m70|MRe5bv+=Jh0XxT%tP$^Q#KNv
zH``U78}+Z2pVPQ>S*LdMJjIY-@HJepuhG+>nK}*>HL}z_Smy+tx`6#r%H+Ds%zHE?
zp|hXM<5tghUo_5CKWoa|&Zvaj=IijyoQs7HlL?>#l?C$_WyvJ#2iTI!HvAVuM&uKz
z1S}bY^hocdC^loKzM%luxy`mQJC+|=V*Golp&?^*W@-=3#j*1dw%fdXd=0>&tYK^b
zx6)Wbx9-)$)_TAJU?SI1&K4`3c<$-f$#t#hJ`D_9y6|E6F>bu?`iNW2;@e5<agv9H
zeb^;B<FV2iAA_Fl!`WX684Gkr)byzfGEfsT`~2FhIf}I~I|#(rtH1@G{y+n5xvDCx
zNctqdh~XR^j!N>RBxBGYrV@1can8@g=gpGk-A@jEH2i}cou`1Jn8v@YA*-?aU%ZP{
zGS@O7bmTQ9z3DuCoIUR(J@i!mk<cZw^AZ)B=(J7@vZZ9riZP3nI!%WD$qRU$oIXJ7
zR5mFO3%@YF2ITq0q8~b=xMS>g7n?4E06m}8kv6hOsTto%4pnZA6nHG@39qW^jMGrA
zwuO=W{oo4@xWVlQdvLvm=qrB7%yhKOzGi)#;ayu=e#pn>w-Jth2p{0~6)!^v5Dx)q
z{1=a{G^;9|u23j-=D2F(VH`TL!rgyuYxE2Nm;;FLc+*JyTsI{`{B;uv#R*<0Y<m%I
zT|p|Hxv|u|qAO8&BGhQEr$=8w$)2o%&9$#0AGQ7~H|?iV{G)@PI~?l{u-?$<hEvN0
zBbVMU7*vQB#!8%UQz(r}q*rEQ+9iZb++P}`BFEU2c>Pr8=^G{7Eu)D;L`H7*;Meg`
zkJJEGa-AJ?ZG5s#cn|$Te%GB^7=+FW{6``m{ubU2KgKU!$?M4CB<^wJHr%(8?-ykm
z>dJX12gj9`dTCDNRfi?_QRI4JD7ziM#+hpIp570G|5_JrGK@9oq=pKE(1)wk{U$vJ
zOtLjiT#uQk8j&VPJZEwkF4Zh?Y-h(FmUG=og`s5d5bbWjGSGa0hbT5M3C0bDgS^r4
zr1V>`@_Y@W8vG$zrX!|unECqQfy5>Pi@N%ibBZN@)4>0cVY{-{o8;t|uV?ZHI3z-P
z?*p|FqGPC*15V(?<nN7w2~j~=e%DB)K6CMX=ECju@_Zd4{uiWpy;A!QQUhtQ;2V%E
zbUhG--T;!<L;D6+{qMo(4Xk<{=msG3zcX16Dpnhi^*?7R)LU<$;%fC@Hf#TTL|Z6l
z?RnDq+Tkrv@v6Sxu=X3>+3Q%lNms`@5D#>a2BNKBJ&k-Bc%UAlH^1_f*eIMoqg%@d
zDncv7O!;<UgGg^LQp23N1YrN7Lj1ngLMl#%bl%+*0R870wQORDrYeAMcIhq+Ofjt7
z5BvU(cSsqwa#n73767QTjUCW%8j1z@PBaqc$AhGanyC8s(xGwJ#vcx5SRd<b4k`Tl
z1-Jwjb0i-b%ln}TJzMhenLHltb*Fx~9B?I0^Y`^GKl$9dAkvM=`w;2%Ltik&2WUst
zL^;PKo<Zt*Ir(QTBLGWknQhMYk0m>GyER&Nx8kr0y#UCop6b)Gsy8=G&>TD8J>pVV
zJ{d+2wv!Xly9h~GoXpbD(?)7|lSzJ4!S=EudO46OrQA6;X{3e&b4j*$7a-Qb%+A23
z*`JWF4pNODb%wj0(1?R%`$ZFZQ~C|PJ$|p=Jom=03VTt#ou7f5-MK6Emm`4tIjq>H
z?BjE9wN<C?tbofS)4`77B6?7fgnr|rvMeV%Po`CO6M_bG^xRHp-hyP8M8Acu7LeWq
zh&5Tis&AfGDN*r{g=j~oZYdu$K70XL*k@RH+!>NR7EO$?!*Jb=Vtt=jcf-Am;*=1k
zXFcm6*Ed6WHY3~m!gk(=Pb24lO@B`*%0M5WI&b@+R9~{A{_==w8PHCv>9<Ex(GR{(
zC#ftt29`YRm=fPKJ%gCOlYEQ2_bf7zdOk5ZFWUmq2kDdCGW`N*Q8gEyLGMhMY{jUa
z!6dv*^Wr${0J?Q`!cG-?K3=OZ!S=B8mD02GccW4;m|JMfHa|@AXGMQfxT=;vr1p~t
z8a*=g=5Vr-jEsMu66;WrjB-Tz`_sz!&Xvda!u+NHi)mCL%)Fis2vsmZ`xU5rL9)%G
ziN-+NQcc*t1d<seb7;BoNupt+sZ`sdX%b(<u?U}jiOE~62bv7!1w9;VwAnM`>w5(|
z5<m7j=f;0fsP|C&|J&}}JGUw0&er<>#V&b)`I_?Gz>|=SOdng|{{<lCr!7vE{OWr1
FKL9NgwTl1%

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher_round.png b/mediapipe/tasks/examples/android/res/mipmap-xxxhdpi/ic_launcher_round.png
new file mode 100644
index 0000000000000000000000000000000000000000..501748acc8fdc38fc10f9aa5f261d106fa15f83c
GIT binary patch
literal 11062
zcmV-6E6LP}P)<h;3K|Lk000e1NJLTq006)M006)U1^@s6Qrv6@001ieNkl<Zc-rk<
z2Y3}l7oC0Sy+G(q2)z?U5kXW?K|xUiC?ZX2f*3jyAS9FoNCK&Z76>VHML~)b1wlZ-
zP^3wb7J5hVr)2)Qv$M(O@e=m+Hv4_^1>?)!w>#&|oO|ca9fbJBFMjchU;N@1zxc&3
ze({T6{NfkC^tmv`0%<DJ99VPIvPpjj_Yn9S<(H*{96;+Nfz}ANBs)$rf^c&1%XDCm
zAb?xG96LaZPyq6@efylI9XjL}bMRoLY5R`^M(;b;Fedj%^XLOdTTjV7^2W5>!*4@_
zf5Y$S|0AD)&(gmke+&0`o*nh?#Yw?0!(0&&DztpL1B%$@SnOvqGCWX9R&|uJtSw5(
z?uq6soq$qv7NFFn>&VYtq#Y&gBKiGo^7A2Sk4Sq;T0Uu*wS4xwNAxq|XD_n9-No*;
zfZiKyP+P<%*e>?FlkZHwdxVO2B)_x+y#yZhD?Qy6B`>W@fcOysWCTjdSxUfpi2VC1
z128VJrDQ8_CEx3*Xp`8c-+_FW)SS*JExQi=j_Q|0VonBrl|0g`0R;XVK$VjHIRRTT
z8_{o-0L*;ySpv@NC$y!gx2&hinufm>@25U<sJB~e#z>~$8NR!wXW}yhzY0c1^^rd9
z2X~aZtR<5&TS&XE8EgQ^oX6r%3fg?uu#p?HBN0XV4jN~T>+~7uvyf+kXM5iB^ULPF
zS%Jy{xuckt=7i>Cw?K1pq6lF2koKPjFo|G;QP?y<*MS&>6v^4e5#m4kZ19Zmtnkbw
zy%T=fsLYXH)OpSRngHe(4WNHoAUiA*@X82LD4o>EK%N<%9p1q+{sNarwkvaBluONi
zlYk(dIj_Zy^4PW-NG2wiiazsI@m-L2BA)@cJhG8!j$GS=V>l;!>$8rlt%Ck}WKeUc
z;oVT@S>`M7=#iC6orFwpl)7{f>EPUFMte~n<l7JMH9FK!G|URQgbagG+HxN*jjYYN
zl`dE$qQKAoQ*FWNF$-5>-qgvUJ}bbxBkxeY;KrS03)T5jr@+sY%+E;2?c=~V6cNa~
zoRj@&k$24f0}CAW=PYZD=HzTt1O5{Zd>x4(N~1y|z`KQa%x42j(E2jqC+1XOE4BZV
zf&VcFv7s8B5v2F}Kl-lWoong~ewldGPs-{+I^}E;;PVmRn9@jk&u8ho=d*!{wSFP+
ztCCJwRy2lhPoYdRhER(}jh4~A(9}o#GLqy;&FV&8d>)rO79)%DXrFLNWF*NWjVchA
zPRU8-a>r6-(b72uYX|bk9KQ@9d9quOmwb?q`W8PM<f7S2UgeU=kdh}Qt3P?M_xY%A
z73mf5-}h0%;sIO|=?i{XdQZuY<)gk8&Ia)B#Edu;9PGeJpbPlp=KGTu`JGynhXc$y
z+LHnjBrINoB4(E6B+v%@$x8xdsSHnOZaxQ<6^2XMpp}@pAH}59<|I(ExRsI>M5fU*
zT<_aTIt7zN2^nWmOlot!XjW2~Gb!^e^0KZm@IT@JvYGaX_Gx0qH58rt4iA$S8Tna`
zdCHPbWaNLq!EfUvNLca!&7AX5L6F04Nl-ZQ3uNT)%!)?wNRK^4h!Qg%l9`|j?R&eO
z2}I2=pyce19Q<}jf`r8n(X`Z$DF*D$F)jdq)*D3@xHv%Uiv+V4JwQ_u-?1+eNZ^lO
z*o?e{D;)fGOak~Bx9}<&7xRjJkf2DMKj##8c)4RtQ~*Y?^Ut8MQ|s9FI)N1GoR;E;
zQnP>O;ODIf31+4pKtsj`*k%G<VL1(fOns}k?BPwV682-1n1NAD>KcL(2Q|#tK)AC2
z{MplG<-Pd|FK_Y{<ELo$B8;Xc&8Ao>iX9tpdQ-A{7pc3+JDs;@(y7VM_=Oln&KhhR
zPO$|3S?Mnk(BI)%oxGJSg|eTB{=nS3Xk7Gb)=dJ*k%>s~CZONXbw6*4!=s5yFq)Qf
z5cLi7FIX+MI^8eoH(2aSH5K@s=PbpQ7JrKy<gCR_mi>qWGrz~a^H*S%w#O4`VXxw%
zLVGohZ7*#puCQozfxTK;KVZKFE3tzSpRZhcp}jg_k365T>=hxwtaOk<=@dhSc(Rg9
zJbXbwNlUsITlWj30q8od+lJ>KI*xz4kcW?4xsH!ry@B^%x{BAII)}&nx)0aQUQ0XF
z5H`VH`yaMad%vE~!-tu@19?~Rrqk!Cy@omK3}(-V*lW9H3$?f7+@%70e-L{c&z#5M
zyAI&S1fWg^um|tcYu<9)j@V1hJ&L!>?B#+z@_doIb8(ZUKN{@$iuDI(&ckT@^j;K8
z@&i4rK>7$MW?m)o5a*3K`ol;Yw0u2YbMh>{dG9{H|KOiu8+7299s95+>0F(`LF_dq
z9l!d-8GPgJy%O!6ynYjp+qoZm>tm0!CQH}hoTI1k_2<|-d*c?dcL4j789|?S+KAX&
zbmSzlSMvKjfAcmTPl#7$;VR?J0!f)f3ME}aVPmR`{y>DI<x3y&QW5EsX>9O+Kt}!C
zylePg2^}SGH%O<Z<{qQdx-RzK{(cJ`1(n*nMMydC;BhLGbg|cZ^(K7q@-=+-{sYDK
zG6;ApFV@Z8tE)ERePq_u#@_N{r-{AqjJG}jQYa#BMZx-jMN1(m>XVq+Rqf<NgUg<v
zW$UQ~xUUe5g`M^QNAEp^UDK9nWwR-f7`rcA*1_I=ob<<0t>sC5qPuqyd8LcJ1&2-$
zd)n#rT3O#yd38@qd$-B^1d@BtTVbs83ox277o+f)-j*E|lBCa+d1c9$K4Hw!-*5g(
z{5_F9+IE)I@^0V7!RvofY|n*&ckN$)>um2DkuTjgY$p;*nLStH?83-;Uq^c|f`7Vk
zyJBbDeeN=xe)xpW_HGk<eYfl~lJn1^LX#6tq29wPTdoum%U^)>i8FS*zvJ3%&(-|T
zYd`;e2|FqnC1JLC>$}Z#q|?XVma`XV#74>Ufx~<0)*Zd<Z9a1WS5PRu-Xza=_Qq{J
z-{(bK#b{C}`w5C$fYHR*ITT}3sTIpx@XB~uwyp80u|jhY4bi8wpuuPkGUNH9@iLrn
z;7`5n0byIL+^FdJV)q?Ztiy|ISMS`TL{jojV}3uR^Xw+ID|hbVPGn#*T2v^RIK)#^
z{z1bgy+N_o-%_GNQoJ9SnUcN6*fIZ##KGICgErRGLbmOGuD$Y$R^x4CbTfp#5!-)P
z6k`3^K=#6S9#Hg7f78LY(bT5w$u<7kKt$X|f;Gp2^#Kdf{wg$U(Z|Mge5D49G7y(G
zM2CSh{5*TGn0EN`HG|ljwg1R-?FB4cg-=|&VGw(12ahRwr{EJA!rtN|CyjleA7qST
zG<;G|iaE2d5Q{0R0tyRrBietLDOP+T8gu-rVE_Ws=6UvD_0v#*oT~tkWz$#TzxCAP
z+5+;9o^nfSD~_Kw{t7W<j7G%nL!CN#F^AZUY15((Q=x>#eM~X>+XDeg5l2~<@UEsd
z<M+chMgVCRn1j*isG*`m9N}!<(SLA+C;3tbOqJ<r>jdzayBx1Pep+u^;6(S{ywlbQ
zAV!4fED(1Dz4w_P#h_WDLS`8K?STLgOCR#{9=c$sk1cpdkk(*x1d#gzqawei7&LiA
zNE-dy^m0WhS-+Vw)!Ehw0If>KcY4d9xP3?H0+$UEK*So%0uyE)K($+XnSK@!SK>?-
z`LQ<wKnvZr`?(2l#kEt{Z_!j~rM9e&0MaZlH5sE}5xpoDO}08*825KX$(d_Sz0t*9
z2mo2DQ}!Ix8k{Gu-K0r^+PqUMB!GxH*ghE^y@g=W*>vjylJ=jT+OptaY7-t$YA*zk
z7R91==i<L_K0hpa@BTk{-=!<KBLTm*iFZ~>0J=Jyh0&CRM<{gkJEn^WiCW76C1fO-
zV~?*r5kN}CfD!%2Q|F3JSGaupE=}`>O+UIMq**5cvPJB{rkT-Eizz0J?gvQPe_U*3
z!oYJP0kSy(j7R`^QeK0cb-2%#op{>bgE)Tw5j<hn0sPUAzu?k(Zc4FU0uby$J)wx$
zOX!P!wZw=}e<x!TzyZZC>R;geN)8DpfKdrx%x1L&kgyk>ic#2vkrbOobB4vxCp0JG
z{H`}wl?P4$P5_P0@VMx01e<OwBBV0j=s#_C5HWn$)Xpy_04G2(O-z-JoOK`d8}d5E
zs8M8)=wLXa+391<S>25jfD=H&EbtUfPs3>V#Ay_x26u)f`O~6>8%oODCK4d)2`2z2
zfMpPXIK#6RVf4+k-6$~7+wd|-l0T8NUNWcr;RN6WC~-{yWKi$WHw`a?#KkX16rVm!
zmiCa(3BU<p83dr!rSoZm+}Gie6srbHkSmP-IHRN`-<u=$&k4W@plDGHqCsOOZ!v5%
zNRmHerd2Pj^u`Hb4g$ccFckHIZJM>o_o|k$hE|$z)UzCkZ4iK95-Q0}jLk#Ay&4)$
z21)WKX3i&ti(gq!IRQ*Z0LblWxoRW*zEED}(B*6N9YF5ScPG!{;oE<wq_ArONSK_G
zh|!?oUy9KnJ#&LFWY@K6<4EPttn_G0INZXX2q3UpolB3NdTt%03%BmzZ+_jcQ*BV&
zAOH)8LKV5N@u?KEfL(nuM1#d$_t0EX!jjGA41d}i0U)WZ`HBsA@5L*MYXm{<(>XfS
zKD7w~Kop3EL(?%DHDwpUtS7~;IDqRa|7pQ1M-(<O&=T~&JrMxx>Qes*D%gP?ZL&gb
zX9j>1W-?)?SRU&9d1Emeq=z5SU)29j^fYhrU^{HC@TWZy0GwSYiLPt4HmKFvb%XNS
zpf*DQF&qkYM4zI*Uv;C{4Gz}X4-gmuoKfuj5teeeojnlX^>2TonqD7U8&027EQM@_
z08|DMj6!y3$ml5)yW$KW=pcV0RBc54<cemeFR%ptZw~|zYl9x4JG}I*4cbm&ZBW}H
z0Fgnl^D!Deeg(xY!El5w>jCVK98vAsUMMMJ2Pc4O2mr?+Z9Qw4+Mr{9-EW@+fQk{L
zBln?7mEe#)oeAJbHB|4xmC21Q!U47i0zhq0Lp>ivgzU@*D0GCwWAf0uooWMy`|IKe
zr=>CdKQ8)Ja^w5vUI3FCoK|?uKmbuvtk(3kL098*Hw{x8bcxN?2GwFgY*Oq4)V)u8
zz;JMc_6uS{dV0Dm`RRhDC3ZE}_Q&Fm@;P9j+L>^O9VdXT5r9R4U`2S~umOPKu(&uk
zZ5?5W0Iq0yYM3btUp@DDqy;PSPiM{>q{BY?@n=o|TO)uN3CgFDpur;`Gt^Zl0t^{4
z#D!+n#m<j4Xa2vY6;g2S^zAGNGb%Z<h!emn2_U9Kd^L7HV7R@AW$g$M6chwLfCrkD
zo?|Hvu>6n=sDP$xM?r_rR^R}00$42p#MJ2F6V?KT8J3k3AXwZ}=|uA>5;A_G1W3+$
zYAFPOSUKqTTRI+;2U(O#&IjWJuu305R4cGWa>S&a1jDdJ(kWP*eZnjR>g-p&x(`ZP
zvR5QPro|8dn)mz_^n6-vNdOTfI6O}2k)@mfR!V?`42(ub9Y&s>WdO@sg+dttT-t?(
zRzOLaM=g;6^uWAD-{M3fWv(fxq70o3XMJ_}VXIcPGdTgQkN}`oU~|kjQ%||HXcGun
z)`kE=gcvt<`8YhhCb`i$OC^A`Sp*J+-hAdfO$UCqXzT-g=Eg00To8m%wY@(lfRz#e
ze1Y(oOCIeyGyyCR(ZUgy2;ka1I<6tP(N#+(fD{>mh|a4kH{!7Ed+7oftb#`EIY`Oz
z!jd)EC2grr_vQq!QUbuw2{UeZb^f#kV3}arRWSh)5)x<#wCBv(LF8t)Eu8?yY@7hr
zNC1Y_iLrM*Kl!{3U|A;u^iE29ncV0>k(+V?Z~|Bh0Vc)$>(k|nwptT_eb~Ks^4wQZ
zYW71;08Rj_BLJwDsQCZ9gL}OTSPp};td$ZVPFy;6|1u@L6}jOfP5@2-t0cgr_{ZMe
z`o0HPCWE@WQUa(X0`y9m*NV&l|KkMU1h7g1M8-ez?$%RI07d%(R7?VSd@(!uHFCp$
zIRQ8UtaAp4d+7afj}CxkCV`X@fPL8Gi}<7#C^?He!JGir=>$g*C-}qeZv&S5_wVnn
zH37O!j(L&X=oTjcCxBHFV8YBho*jeV)S3XjdiA0bsPp*ojmga}Tgy5CB+0?iqOh&7
zZ2Br(YQakEF?YF6Cns_O*uX;A`03ZZ-tG7rU^zHA*iCB!ygO)UZIqOG*3t<8z=sM5
zGxr_F+s|F1fc^W%ZM;A4D(#feZF_LFjPGcvK~4ZWN`)8~ebN23H-fY#K>POXVIj<;
zX@?F~QR0$6EtLQ;vcukhWyenAOSkV9TS^2nM*kpW2>p2v_9Qa`Cx9IgAaMy)hWN|5
zL9@Dm<u+~FxGE+<!DbkQ0?_P5dn}OvFq*esy$SESpuBwG^}F}*ls&mrCUFATdOGxC
zjD|-XA{ed+Sgv2czO%Lj0H>ieieLDXr4azY-+1XdoSSz|r`=$a2?>-Y$8iE!;f^r6
zAtoK8u<)I95GPn>m{vppSS@oTjPaxdh@F>ZX*<GT1c%z0cXjNzzX`xPt=+~6V9NxE
zTYyo>m>&Sc49kv+Qz78PEnBwqB0o!^X~{D!X;VD3x4zp<<%~X^wf@vOMNTm%fR%2F
zkC}r}|FDIC;o2gWwIYB+_wL<cFQq>kA3MU5%Az<E=`#P&af5VN{p_`z0Jd5dWoi;e
zy+WWsu7BI$U>|nQg>nJ}2M5En;fID#>W&KI=w>GYR6K@Iry)A*%PqS&0c^D}R@AI~
z)U|ggV7O(cPF`9Q;O)2HhA5CP>M`hb!m7KLQW_DeIUK%n-5?z{ZSO%&09!4MI4<TM
zdZ*(DfZ_UW+IVV9096Y(6&@iU)bfqmC^6%Vr9^@ZSsYhMZ6G*}(Uiq%IROl`A%1k!
zMbxAT><9J<3JQ|aAQaaIQUwMEy0Drm<x%{?ZEOaxyXw%epLqWz!_*?3y5}G#fT_+2
z8~V-f1iMuL!_}%)Q@jXPK!Ad}KnRsZ(JAxo^?U$0b#cwfvwGWt<Bqy-+QA863IfbX
zgTlBgDR%w*d}x((9SGo5y?PM*aTydoW27t+WLIZ|IgsZbvEz5WZ9!?4nhMVH<pj`6
z|L0STLS^XB2Siisii5b5wgjMS0kvz>_-Sd>Z&;^-ondBngiTEVXf-m|=yhfnEZBkm
zhr{nU0gUVb!$3V@#v}Azm!1^6fq~xItO2O422=0ZAGLa`5lUQg+7cXLa}WU9m|yp6
zEWD+p9VOqlZuSp~>~I3;xdIGHa3doxpe8T9O0f$^eCbF4hf0+yxzSK4BoxIi*kB2c
zu(=2TzChZ+WAq$eZEXB?{U-io{V$4Bu{i<s)c-UEVl?CcRi)UiP{CcB5NMH5C_N!Y
zJZP*uiinT41V6xB1dtZtfCNeb135<kerGLek#Yj)ssE=YW7I2jF2$^fT^DUbphY4<
zs7gaW>P-(C@fk|Vd2DV!z?=k-oZ@a9w$ogqD|hZ`(EI=3<JWHBq(6?*H3V%p;cx=z
zxd1jH_6ho+O9;hmrTYH5M1YD!f|wH#fL?9W1SMve-48G~0VEQ@Yyi8^M(@~19o)r-
zPvY2phjIU{zu`tp*J00j%XPXpCxCt>(S?zdE};5NT2jpF7y&Bc2vZH^<yDIOyE2-d
zy4u{S(Uw2}V>V6zJsjW}bFnaZ_-3mA0kfk1chp7x6ga~4pu7sTN|T=z(5R@9MUsv<
z0XP9HV+Ift;38*X)U8i6#cH|00E%559AI@v*cnx;29cn0=)-Pp$TV}!oHB?LfD@p&
z{2yV=)LW?a8=WXtt5orm>3^LFQ0NB)1bDKoWz|r0>LzoBLpcFB0W^j`XU-$?XXGw|
z(Ha!1B4%~g|3&=(I4-qZxd7@1jEwxIVAFvaWDq9+Cx8a|laC_fG5WO6Oo~wvt8V(}
ze|0}VS`Vmz-u<u*nzhJmr^avsZ~|y5fg+5Gyop-0d7ommV#QJ{{_mhK0Seax{QbRY
zauGr=peae;7pcI-3BU<p4gyHgpXq5>7&v?zlRss2%>aF#401<hUMNd*tOt!8LN^_l
zVV4*u04Kn+i(fM^`slOq6q^Cn%CV9@j)oZhi-tqJy+s)mSg$E@I*yy8Xf7uJCxCkW
zA0ce~SyZjs%M_d9nw6`e<WEr<gpe01UAY34L1U&SvRT0H2;eq%8Ge2BCLFW(5Z-k@
z4_~@{hn8PFdF>|MJkxLMZ#XdXd!35%asp^s{v>KPMqdnCLNO^~)60<Z=UEw4wrm;d
zX?N)SE{a=t$J8>&oCE+sS6{M*o=^`PT#CEzO+tn}7cbMa2yIg$I03XseE{co^yGV}
z)mxuXOqMBA*3k0j*=UfjuP<_U2c`w0iLu|(Ss*#*F`JTXngBt|*W(@MF466cO70J{
zLey^UYLRjRD3w11bJLQs&_C=aibcRAVAFunAI*zm2zjD_fO2$EtXtn6Xx73frkn-L
zMSzA&*WsP#^R%`BB!IH*P6|_^IRO-o{)7a=H&dRXcRv_Ru~@ovd5TQ~EPg$IQA~7(
zE0I<UO^W|n^aV__GHeb40C*P@FzE`Ei?{CJ_CIV@lp4(mpm60kW)2nxjo3>tSchU!
zbbj58A%C9D4FX4_R3#`oT^;ob9f;x=nk*!04g$Qoaw8qd^|7_?ufMUk0)K!LK;h^g
z9)r>QT}M(3`d6%^=KLCk07ac)FR#+n@c;%!#_ytgF-<TFn3@1kMji6}^|f{F?mgUm
z#Rg6QLyZ2g@HKeE9|U{#C<Z+}1B~VT7Viv0)j0nOmFenR_W^@Z?EHLFq)D5a0A&`g
z!kf;VH;BEEZF@KY^v(OB+CMrUz1L+l#hzb<Din)GbbgC-hCLYb;0x46<EL+h9XWIs
zFk(c=)C7Qd{-Mi;sYM#U{|G05zN3G13Ksee{f&+ORTu_6jOqLqzcLKF?*08M(J)l!
z&%Z#^=Ni3sP!mFcV+JC?ECqlxeQnTH^ngD@*jucS0Q&kjQd@THlyS8`<ou70{2yx9
zX&}X%zkg*`wZzp_qyMuZQDC5}Yc=W%jGnxb)@77V=`+T;z1ctD+{@Pt(qR$5{h_Ef
z=z%=L)CP^-d+53LN)yR-*g!MRYy|{|cD95?J#AW=g6h|L5nyrx7J7$npjZRUiKD+f
z^l35zNY1c)7N~*Ry!9@cns`-Q95YUikl%upc++XaX!6cK{%mozL1E>1^@%fj+k<L!
zO3pukW{BqldqWiXIC|y$d>wuRz0~p}inXGnzoRL^uO1Q=&l2+~Tb=x@jY3An!|oh$
z!%UXJ9b%}+0IRF|b{Ld2fyxfgv*$?cjs11M-u6J(gPc<Gd?R=4!?*R+<)^OS!Y?gf
zujri){bjG-e!yvBuhpte#@p}$8J}Hx&7~OgDN}=DPQ;$GsYicJzJL#MbE`<1>EY2B
zjgL8iW-T&WiUjxp=k;(D@3HAxY0VyRgy9@weeBIWcwA9!P-wLXQJ|unK7d?s=mhps
zn1=+!JgBFv%sTp)VtZPrJiwyYz~O%)clRKQF*mnLEb?P=?XPJTkj@k?RWJ*D)U6kq
zmi&lHAtN}%P-!iE=K-2@rb}o2e(4Hsn6pl+V_Co1xgTHGu{P+RSFY0oy|u9yMn+{_
zYmtJcua~`6t0cdHTYuBp9*q9YSFAUV^r5wWA|Ii49YVy>pJJ?(TI9zp1bB8`0CwZM
zxYVFw(XYlXf;uhK7cjysP$qp9UViMf*0nkzi!*rrFIuN6gS`wL(jm^?xQ)N~=~t~Q
z#DOymS*P0CgK~_Jg{rMR@O4shk7*qdf~wgAh&?AQ%0DWWK0ttpGqI{i|K$`@E-rx-
zV<wLKD~$-b(`vkKZm=#;6Z!ZwL*da!#54)xtPNCN{4JFhO3w?1l7-;->e^Z0Ge9U)
zsXjkNWK5S0zvyfaKneDgj?AF@fo!_kgRIr0KmNqm?kGEg7GxTL-_g+EFUa>vg?yir
z$lJR)#gwaSRf;hYYbJ~PE1d<LSw2P?<Q))5LscDvd!eX=hcr#XNKqm0c`I<=ExYix
zzb_KU_O8a0;bEgnO72nIZ26DW$<)UlL}_|&-ibG#xqvTgum@y3LF~a%t}hZNR9}0b
zt08){@$`B6eD^fi10XK>^CV5z))$Om4+{Fi>Nog3n(Tq($tIfrB_eyAi~)WseHLS3
z*u?+Po9zctEO`giWP8%3oF9uA{h!Sg6+=_yk(XC3SglhH9FYjCcWeW_QPLzp&QxCf
z9qzPl8;;zaOBaBa5qXsS$59-zZ8v^-#Rlv?cbP%$Rao>bwKws%Kky=I?-V83S3mEi
z_B;*f5Q05eAp3yW3*U91z#hyFV6Vl>jRti<MSH99`)jvS-)Awi2ea*0Kkvb<R&O%A
z!>cBJAlXgzY3~Je#aF~qd4?@FbA&y$cZkL53ch$FG;-2*>JOwCH%)>!D{qqYiJFat
zK3_qqb7P7jaiP^>9A704u{SMG?d>f}p@uKDMd2~$sXt%=Cr$EZ!Qh{fhJ}z(mr$*S
z9YpCvF;tde%Ei*8k0ys$E`=%}uhO-t6#BUPKr|`tU(j}zaNaO)76yL0+Peq~!zVmc
zweK`sls*(Y71X4U3R&>=tP~P=>%gv^2CCknF?1I{IK>ul>O5~21b$lFb_!N~(jx(|
z!Z1_G8l+E2Qb>#nfrfK-h6ieh293y~$5<w3nf=IB-YlT88nfW^3f=pyrWgUNxVpX|
zi~3l+^eHJ>PFf%fq>CBg*bE)Fg`VY<1X3uAlfZtPxY!jDOPt=&UkNsvXjt%awZy28
zHc|){E&N#3F6s|Bm96*+8a8o1EOc<E*q%p!7V-IV=wZiz3a_z{r=<J+)E2y~iU682
zMwvfQ3Zzhd$A(SYq7jo$!v0Q90y_o&)Fdnn7=Ff~UcL5!3FKL-u1xovHPh2r{(!iJ
z($gQdvo&yN-s(d%YSIM=eW9f6#~c{;;`FkjR)a?5IlTDtC)DW`F;RtK#42^a5`RD{
zv0fGV`85EC$KmyM!Dv)e9w&jlGy;Cu*BdhGio+|dyHlsv&#xg854JADFyd+rAx~|k
zkhDJF!xlEG$w=@ikwbZq_s!*yjmsVw`3H}@>hS98J!IgoCUbh_@qPi>$R&4*rA0D$
z;1`f5BSGu8x(XvAE{J(UnNN6#)W)Jc0RAb7qU@1IezEUgU|5hAyPRwd{3S((%F`Dh
zjslGwf?9kigpEH**9j7rJmn;?h4hA;e3}RuIg2>F!%jOirJBDH>GMVm{N-&S;-g=9
z6w2F)FQk!Ut-5Vd=-B;q#W-n6K3_4m@)cef#3s(fLf@|rI##RkE(O1@Z)4i$MZ=xW
zwhn$x;ZYB^!9gT}hd2{B`un#)gNAQK6Js&W9ZKbXfmJ59LH+)5(OBp?c!#5(UrP#p
z4?56E;P+=3up97e+C(EKL4i-;<mA*0^$E>J;nQ)!t}+fT>v49WjLE1dEPU2?CGAVl
z{Ne!jf(GzA*c13mB7x`=czQ}CXrlVG&vaqTlz(LlX525ZT*;GPP`W&9{6p1;pT<KU
ziVm-51DWQxd+--W0?9Fk<hgpJi6rRIIfT5>bE036nQxBcS$R{a_aoz>a`OdM+xNbv
z;3s`Z`%(;bN)9gve+iDU6e6vPTwH1+58uY*MK*S<Q|BFFz_1-^l8Da+mNM!?F8|o6
zSm-@?mqV?(9iY!h-!)-bz*S|TPAS^M!CwLiq{T9^!i4BJH6kyysj6Gw6k$~4KLy#O
zd^Rxes1GMK!s1p~_&=(Td(35eUvhS9vOX-!`qVB$orYR0^I>@*Py;d$d3ZFGNZ`==
z?LI=ts3Wvo9^Gb{^O(;DCLZ<SozGrGPihgq9D3BDRoi~hHzN3HUj?$h3y`hwa`0=B
zHsQ*WqM!z(cpQ_nbA2ei2-?!g&8@kr$AHDc$jE;ot5B>$m(6Db;{w0%9ZyZBj;_OJ
zy|YMPyh8Opl*NQT62qJoS>Ji7<@Rv!Ye51RmKN(@;2cAh0Ol9a+2GB0z7z(HI)LEJ
zC@9LC%w>_$4y6Y0$1M<J`~AMoRkeC^AoM|j-WRKYRFg$|JXs%d@asqdd4$N9X-QEM
zIXgEfm<^nqUQ~7Ml^_fqcLjDA(ZamonDH6GKsukiOE~=pc5a1^zM}f5dop?F&1Iv0
z1DWJ0%lb+l>EYnlGfXP^1>Q1A1Tp2>MP`FWjXI&816K<pCOx1bQ|L621e^%kfW9!S
zNoG?bEg^|**n|f{_kL>}>NfmPHtIK~eIWLQBze4LdY?ynjNlh|vRHCC7GVgay3^ga
zDVq(N5huAXam=;}qbEP2iFNesGCm_H20pwi06ysak&%yto*}<Dy!J+ac(-)aC-0hR
z{Xo_S<yfD1a=mXFNhF0z<+B0t4Llklkpz?os<s`!7Wxg{EsTzO49Dw;c||fJ@G5oB
z0w012uvKwX)MKGf=pNNu?++^gz9@MbF|98j^?hUsZCvu0C{!w+4J6;71{s;^Qs2PC
zqY0a;gD4Tc9I{;)8SxNB#Ic36octp4$ezfcE_2{r%qD&lq%4n^_)zE*vV#FXs8H@u
ztzVbE<3g=3AN6^N(~PqL=+06s`cRe0qIwXSko*HGjas$)Qs_D8d*SPGH_-U$Slou0
zDfbidZI%dypf5^%h@4HBfyq0$DfAfdgF~w~`>_#U1imPD>M^-fmA&Iq?49#b-;5+s
z50^@jq5w8ESEDPDbnQTm2o0KbR(0#0CJY{VTo^m$DRr3Xf?Bp*BIV0KV;i7{XM=L>
zFxrD-J7(GwVZiVcsxDv5b7;`Ga}nT+X$<w*YHl_5P6K2S9`*1h2fvwT19e$cM&=*X
zELcQycW){Y!O_dBrK(-0Q9`f5>xHkr$%9o*x~K+WRIzH=(`RM}OD2S-Q68QFKo9;8
zY&V6KTX@!<gVw9sbsP=vpaA&f-N+WRYRYte84YrWkNOrGDRO3EI4_xhP^n->fXZXg
z$?iVQn6oU-3N`Dzuj<%!vd}wpO92_cagJLc$|Uiii>DfA2y>94sO|hB-wnPi{2tD1
zhMXw?{m^lFLa&f5s`tB0bqK88NjBOyfp<aQiCDQ$8u2U1fbYfLsWXr8SdT1HCxSF1
zRAqX$HW}#}(eP=Z!`zILp?3XFs`o#M5WX0&P6!$GmoO&kAK2ChNncbdC2JrAU471a
zmKf?yTW}Zo+bnU5sJPOt0q|tzf@*uj(HF)}{)c{d@(il?K8kdxUGD?t)QZ3-&&>3F
zBNoe-G=4<|z~|t%Vj=`EM-JlG)oDl%oIL9Axj#b&$I|86z>>e}lO9P#Ds85q8Zq%6
zP0Nm)h3Q5eAO)>>oh=drqAH^FkV~ki1w<*SsefL&k9be<{y;zgJ2>$$OMw$RGV-1f
zI{Gi6PsnD~C!Z&i?>vZn_cx^HC<Hx`Iq<Ba%&EoRK^68+0yyw(p9r2TZe5Bk;7IaF
znicAcWFQjBw_tHiB1Dj5>C$Z+UTM_}Mpf0PeX@nVU;QEs9(6((KH-Kidh)+CLp3s<
zI+>IlvGYZOfa95%A<8LDvjP2Grbr~v>mXl1e!<B)F!KP}U~~D{ssB=2A)`-#ZTJp!
zgn#+9-tgTZEkrJF;2EgI5kX&1Hrh)vryP41KJ49a;M+J6BxhL)pUP*2%0<Wknpl!a
zK1F<nrm~UsMMposR#Y~%ZWltNQ<N(B^Av~A`>iCJd$TYwY!4yG5jr}59e$30_o6U-
z(iJjdUl&G20cmc~e~%=;8y;~HcD@Ni#{5P8<}m%+-l0E{d#!+bs^0r}GS~n)<LK+#
zS_bxJvhURdzN<{$)Kvq0Wwp^>3TH~rEeF0G5W%6CS;3#_-}17Vp$2md>(Hbx02PxG
zbcQI5D#=M9wIBx4=8j%otz0Tse#bqqPDi);jXrj3*d*Ak*-KsBn!VDM{+s*`{vUh>
zKFgBLUX<zfX3v0~eFr2)Z<^5WCMKrUVc)riY@{#Gp3T3Q(cXar-~OUT@>#(}0~rFC
zM5xGgvYZffhHxg5iL{0u1b`mCzD-J(FaKis%9UUA4G3uJLq=lvq6E>Z!Qa9?;9hV~
zxHs4UTVT^!)bY{?F9&=T_MOYIXHW;di`r<<qdI(6D3}%GWbj}T!I$Y|i4YZ-l&H#L
z<u#bZsKvlohXJ!50cit9kjBt_eSM|YRISn9OP^x^hta$qyH_1{?^?`O4Q9J4`z{q^
zfG@?qvp0JNb<p$Co?nW|;4JeQBtl3tL}?}~$}x#ik%6%a17<Y_(&{pjNUf$?lmEXO
z>~CIR_o&M5S((`=&uo@q-$5GTCBRFX-kF1*U;2_k6CtD-!kx(qF9twg=2(_uzzkp@
zEz1Ba5k+bhSd;%w`dnG|w*l;4rP#fFnJq77+g&!oYXUq6J--Y>26aMcB84jhqnm8T
z@RX6nTQ-AeYSRDrlKq{B4CHRih9<CU0z3yjzf497bz(^5aFUV4Sw<F3P5v4A-#Qip
wbPjfYSsWpXBZsOOvS<Ressxa8z#H)TKM#U(IgVDiO#lD@07*qoM6N<$g848ArvLx|

literal 0
HcmV?d00001

diff --git a/mediapipe/tasks/examples/android/res/values/colors.xml b/mediapipe/tasks/examples/android/res/values/colors.xml
new file mode 100644
index 000000000..69b22338c
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/values/colors.xml
@@ -0,0 +1,6 @@
+<?xml version="1.0" encoding="utf-8"?>
+<resources>
+    <color name="colorPrimary">#008577</color>
+    <color name="colorPrimaryDark">#00574B</color>
+    <color name="colorAccent">#D81B60</color>
+</resources>
diff --git a/mediapipe/tasks/examples/android/res/values/strings.xml b/mediapipe/tasks/examples/android/res/values/strings.xml
new file mode 100644
index 000000000..1e6b68a9f
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/values/strings.xml
@@ -0,0 +1,6 @@
+<resources>
+    <string name="load_picture" translatable="false">Load Picture</string>
+    <string name="load_video" translatable="false">Load Video</string>
+    <string name="start_camera" translatable="false">Start Camera</string>
+    <string name="instruction" translatable="false">Please press any button above to start</string>
+</resources>
diff --git a/mediapipe/tasks/examples/android/res/values/styles.xml b/mediapipe/tasks/examples/android/res/values/styles.xml
new file mode 100644
index 000000000..5885930df
--- /dev/null
+++ b/mediapipe/tasks/examples/android/res/values/styles.xml
@@ -0,0 +1,11 @@
+<resources>
+
+    <!-- Base application theme. -->
+    <style name="AppTheme" parent="Theme.AppCompat.Light.DarkActionBar">
+        <!-- Customize your theme here. -->
+        <item name="colorPrimary">@color/colorPrimary</item>
+        <item name="colorPrimaryDark">@color/colorPrimaryDark</item>
+        <item name="colorAccent">@color/colorAccent</item>
+    </style>
+
+</resources>

From 38baaa00b1653e4f5dc9644426bb24344a7be6b1 Mon Sep 17 00:00:00 2001
From: MediaPipe Team <mediapipe-team@google.com>
Date: Tue, 4 Oct 2022 10:01:17 -0700
Subject: [PATCH 3/5]     Fix GL Tensor (backed by AHardwaBuffer) GL to CPU
 synchronization.

    glFinish() must be called on a valid gl context

PiperOrigin-RevId: 478816905
---
 mediapipe/framework/formats/tensor_ahwb.cc | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/mediapipe/framework/formats/tensor_ahwb.cc b/mediapipe/framework/formats/tensor_ahwb.cc
index c839cf5a2..b11f6b55b 100644
--- a/mediapipe/framework/formats/tensor_ahwb.cc
+++ b/mediapipe/framework/formats/tensor_ahwb.cc
@@ -371,7 +371,7 @@ void* Tensor::MapAhwbToCpuRead() const {
         if ((valid_ & kValidOpenGlBuffer) && ssbo_written_ == -1) {
           // EGLSync is failed. Use another synchronization method.
           // TODO: Use tflite::gpu::GlBufferSync and GlActiveSync.
-          glFinish();
+          gl_context_->Run([]() { glFinish(); });
         } else if (valid_ & kValidAHardwareBuffer) {
           CHECK(ahwb_written_) << "Ahwb-to-Cpu synchronization requires the "
                                   "completion function to be set";

From fb3b0e788e16106ffa184a36aafa13a1030c1376 Mon Sep 17 00:00:00 2001
From: MediaPipe Team <mediapipe-team@google.com>
Date: Tue, 4 Oct 2022 11:09:04 -0700
Subject: [PATCH 4/5] Open sourcing model_maker/core/util and
 model_maker/core/data

PiperOrigin-RevId: 478835650
---
 Dockerfile                                    |   2 +-
 mediapipe/model_maker/BUILD                   |  22 ++
 mediapipe/model_maker/__init__.py             |  13 +
 mediapipe/model_maker/python/BUILD            |  22 ++
 mediapipe/model_maker/python/__init__.py      |  13 +
 mediapipe/model_maker/python/core/BUILD       |  19 ++
 mediapipe/model_maker/python/core/__init__.py |  13 +
 mediapipe/model_maker/python/core/data/BUILD  |  68 +++++
 .../model_maker/python/core/data/__init__.py  |  13 +
 .../core/data/classification_dataset.py       |  47 ++++
 .../core/data/classification_dataset_test.py  |  68 +++++
 .../model_maker/python/core/data/data_util.py |  35 +++
 .../python/core/data/data_util_test.py        |  44 ++++
 .../model_maker/python/core/data/dataset.py   | 164 ++++++++++++
 .../python/core/data/dataset_test.py          |  78 ++++++
 .../python/core/data/testdata/BUILD           |  30 +++
 mediapipe/model_maker/python/core/utils/BUILD | 100 ++++++++
 .../model_maker/python/core/utils/__init__.py |  13 +
 .../python/core/utils/image_preprocessing.py  | 228 +++++++++++++++++
 .../core/utils/image_preprocessing_test.py    |  85 ++++++
 .../python/core/utils/loss_functions.py       | 105 ++++++++
 .../python/core/utils/loss_functions_test.py  | 103 ++++++++
 .../python/core/utils/model_util.py           | 241 ++++++++++++++++++
 .../python/core/utils/model_util_test.py      | 137 ++++++++++
 .../python/core/utils/quantization.py         | 213 ++++++++++++++++
 .../python/core/utils/quantization_test.py    | 108 ++++++++
 .../python/core/utils/test_util.py            |  76 ++++++
 mediapipe/model_maker/requirements.txt        |   4 +
 third_party/external_files.bzl                |   6 +
 29 files changed, 2069 insertions(+), 1 deletion(-)
 create mode 100644 mediapipe/model_maker/BUILD
 create mode 100644 mediapipe/model_maker/__init__.py
 create mode 100644 mediapipe/model_maker/python/BUILD
 create mode 100644 mediapipe/model_maker/python/__init__.py
 create mode 100644 mediapipe/model_maker/python/core/BUILD
 create mode 100644 mediapipe/model_maker/python/core/__init__.py
 create mode 100644 mediapipe/model_maker/python/core/data/BUILD
 create mode 100644 mediapipe/model_maker/python/core/data/__init__.py
 create mode 100644 mediapipe/model_maker/python/core/data/classification_dataset.py
 create mode 100644 mediapipe/model_maker/python/core/data/classification_dataset_test.py
 create mode 100644 mediapipe/model_maker/python/core/data/data_util.py
 create mode 100644 mediapipe/model_maker/python/core/data/data_util_test.py
 create mode 100644 mediapipe/model_maker/python/core/data/dataset.py
 create mode 100644 mediapipe/model_maker/python/core/data/dataset_test.py
 create mode 100644 mediapipe/model_maker/python/core/data/testdata/BUILD
 create mode 100644 mediapipe/model_maker/python/core/utils/BUILD
 create mode 100644 mediapipe/model_maker/python/core/utils/__init__.py
 create mode 100644 mediapipe/model_maker/python/core/utils/image_preprocessing.py
 create mode 100644 mediapipe/model_maker/python/core/utils/image_preprocessing_test.py
 create mode 100644 mediapipe/model_maker/python/core/utils/loss_functions.py
 create mode 100644 mediapipe/model_maker/python/core/utils/loss_functions_test.py
 create mode 100644 mediapipe/model_maker/python/core/utils/model_util.py
 create mode 100644 mediapipe/model_maker/python/core/utils/model_util_test.py
 create mode 100644 mediapipe/model_maker/python/core/utils/quantization.py
 create mode 100644 mediapipe/model_maker/python/core/utils/quantization_test.py
 create mode 100644 mediapipe/model_maker/python/core/utils/test_util.py
 create mode 100644 mediapipe/model_maker/requirements.txt

diff --git a/Dockerfile b/Dockerfile
index 462dacbd4..4d6c68e7e 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -53,7 +53,7 @@ RUN pip3 install wheel
 RUN pip3 install future
 RUN pip3 install absl-py numpy opencv-contrib-python protobuf==3.20.1
 RUN pip3 install six==1.14.0
-RUN pip3 install tensorflow==2.2.0
+RUN pip3 install tensorflow
 RUN pip3 install tf_slim
 
 RUN ln -s /usr/bin/python3 /usr/bin/python
diff --git a/mediapipe/model_maker/BUILD b/mediapipe/model_maker/BUILD
new file mode 100644
index 000000000..cb312072f
--- /dev/null
+++ b/mediapipe/model_maker/BUILD
@@ -0,0 +1,22 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+package(default_visibility = ["//visibility:public"])
+
+package_group(
+    name = "internal",
+    packages = [
+        "//mediapipe/model_maker/...",
+    ],
+)
diff --git a/mediapipe/model_maker/__init__.py b/mediapipe/model_maker/__init__.py
new file mode 100644
index 000000000..7ca2f9216
--- /dev/null
+++ b/mediapipe/model_maker/__init__.py
@@ -0,0 +1,13 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/mediapipe/model_maker/python/BUILD b/mediapipe/model_maker/python/BUILD
new file mode 100644
index 000000000..cb312072f
--- /dev/null
+++ b/mediapipe/model_maker/python/BUILD
@@ -0,0 +1,22 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+package(default_visibility = ["//visibility:public"])
+
+package_group(
+    name = "internal",
+    packages = [
+        "//mediapipe/model_maker/...",
+    ],
+)
diff --git a/mediapipe/model_maker/python/__init__.py b/mediapipe/model_maker/python/__init__.py
new file mode 100644
index 000000000..7ca2f9216
--- /dev/null
+++ b/mediapipe/model_maker/python/__init__.py
@@ -0,0 +1,13 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/mediapipe/model_maker/python/core/BUILD b/mediapipe/model_maker/python/core/BUILD
new file mode 100644
index 000000000..10aef8c33
--- /dev/null
+++ b/mediapipe/model_maker/python/core/BUILD
@@ -0,0 +1,19 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+package(
+    default_visibility = ["//mediapipe:__subpackages__"],
+)
+
+licenses(["notice"])
diff --git a/mediapipe/model_maker/python/core/__init__.py b/mediapipe/model_maker/python/core/__init__.py
new file mode 100644
index 000000000..7ca2f9216
--- /dev/null
+++ b/mediapipe/model_maker/python/core/__init__.py
@@ -0,0 +1,13 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/mediapipe/model_maker/python/core/data/BUILD b/mediapipe/model_maker/python/core/data/BUILD
new file mode 100644
index 000000000..c4c659d56
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/BUILD
@@ -0,0 +1,68 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Placeholder for internal Python strict library and test compatibility macro.
+
+licenses(["notice"])
+
+package(
+    default_visibility = ["//mediapipe:__subpackages__"],
+)
+
+py_library(
+    name = "data_util",
+    srcs = ["data_util.py"],
+    srcs_version = "PY3",
+)
+
+py_test(
+    name = "data_util_test",
+    srcs = ["data_util_test.py"],
+    data = ["//mediapipe/model_maker/python/core/data/testdata"],
+    python_version = "PY3",
+    srcs_version = "PY3",
+    deps = [":data_util"],
+)
+
+py_library(
+    name = "dataset",
+    srcs = ["dataset.py"],
+    srcs_version = "PY3",
+)
+
+py_test(
+    name = "dataset_test",
+    srcs = ["dataset_test.py"],
+    python_version = "PY3",
+    srcs_version = "PY3",
+    deps = [
+        ":dataset",
+        "//mediapipe/model_maker/python/core/utils:test_util",
+    ],
+)
+
+py_library(
+    name = "classification_dataset",
+    srcs = ["classification_dataset.py"],
+    srcs_version = "PY3",
+    deps = [":dataset"],
+)
+
+py_test(
+    name = "classification_dataset_test",
+    srcs = ["classification_dataset_test.py"],
+    python_version = "PY3",
+    srcs_version = "PY3",
+    deps = [":classification_dataset"],
+)
diff --git a/mediapipe/model_maker/python/core/data/__init__.py b/mediapipe/model_maker/python/core/data/__init__.py
new file mode 100644
index 000000000..7ca2f9216
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/__init__.py
@@ -0,0 +1,13 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/mediapipe/model_maker/python/core/data/classification_dataset.py b/mediapipe/model_maker/python/core/data/classification_dataset.py
new file mode 100644
index 000000000..9075e46eb
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/classification_dataset.py
@@ -0,0 +1,47 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Common classification dataset library."""
+
+from typing import Any, Tuple
+
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.data import dataset as ds
+
+
+class ClassificationDataset(ds.Dataset):
+  """DataLoader for classification models."""
+
+  def __init__(self, dataset: tf.data.Dataset, size: int, index_to_label: Any):
+    super().__init__(dataset, size)
+    self.index_to_label = index_to_label
+
+  @property
+  def num_classes(self: ds._DatasetT) -> int:
+    return len(self.index_to_label)
+
+  def split(self: ds._DatasetT,
+            fraction: float) -> Tuple[ds._DatasetT, ds._DatasetT]:
+    """Splits dataset into two sub-datasets with the given fraction.
+
+    Primarily used for splitting the data set into training and testing sets.
+
+    Args:
+      fraction: float, demonstrates the fraction of the first returned
+        subdataset in the original data.
+
+    Returns:
+      The splitted two sub datasets.
+    """
+    return self._split(fraction, self.index_to_label)
diff --git a/mediapipe/model_maker/python/core/data/classification_dataset_test.py b/mediapipe/model_maker/python/core/data/classification_dataset_test.py
new file mode 100644
index 000000000..f8688ab14
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/classification_dataset_test.py
@@ -0,0 +1,68 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Dependency imports
+
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.data import classification_dataset
+
+
+class ClassificationDataLoaderTest(tf.test.TestCase):
+
+  def test_split(self):
+
+    class MagicClassificationDataLoader(
+        classification_dataset.ClassificationDataset):
+
+      def __init__(self, dataset, size, index_to_label, value):
+        super(MagicClassificationDataLoader,
+              self).__init__(dataset, size, index_to_label)
+        self.value = value
+
+      def split(self, fraction):
+        return self._split(fraction, self.index_to_label, self.value)
+
+    # Some dummy inputs.
+    magic_value = 42
+    num_classes = 2
+    index_to_label = (False, True)
+
+    # Create data loader from sample data.
+    ds = tf.data.Dataset.from_tensor_slices([[0, 1], [1, 1], [0, 0], [1, 0]])
+    data = MagicClassificationDataLoader(ds, len(ds), index_to_label,
+                                         magic_value)
+
+    # Train/Test data split.
+    fraction = .25
+    train_data, test_data = data.split(fraction)
+
+    # `split` should return instances of child DataLoader.
+    self.assertIsInstance(train_data, MagicClassificationDataLoader)
+    self.assertIsInstance(test_data, MagicClassificationDataLoader)
+
+    # Make sure number of entries are right.
+    self.assertEqual(len(train_data.gen_tf_dataset()), len(train_data))
+    self.assertLen(train_data, fraction * len(ds))
+    self.assertLen(test_data, len(ds) - len(train_data))
+
+    # Make sure attributes propagated correctly.
+    self.assertEqual(train_data.num_classes, num_classes)
+    self.assertEqual(test_data.index_to_label, index_to_label)
+    self.assertEqual(train_data.value, magic_value)
+    self.assertEqual(test_data.value, magic_value)
+
+
+if __name__ == '__main__':
+  tf.test.main()
diff --git a/mediapipe/model_maker/python/core/data/data_util.py b/mediapipe/model_maker/python/core/data/data_util.py
new file mode 100644
index 000000000..8c6b9145f
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/data_util.py
@@ -0,0 +1,35 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+"""Data utility library."""
+
+import cv2
+import numpy as np
+import tensorflow as tf
+
+
+def load_image(path: str) -> np.ndarray:
+  """Loads an image as an RGB numpy array.
+
+  Args:
+    path: input image file absolute path.
+
+  Returns:
+    An RGB image in numpy.ndarray.
+  """
+  tf.compat.v1.logging.info('Loading RGB image %s', path)
+  # TODO Replace the OpenCV image load and conversion library by
+  # MediaPipe image utility library once it is ready.
+  image = cv2.imread(path)
+  return cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
diff --git a/mediapipe/model_maker/python/core/data/data_util_test.py b/mediapipe/model_maker/python/core/data/data_util_test.py
new file mode 100644
index 000000000..56ac832c3
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/data_util_test.py
@@ -0,0 +1,44 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import os
+
+# Dependency imports
+
+from absl import flags
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.data import data_util
+
+_WORKSPACE = "mediapipe"
+_TEST_DATA_DIR = os.path.join(
+    _WORKSPACE, 'mediapipe/model_maker/python/core/data/testdata')
+
+FLAGS = flags.FLAGS
+
+
+class DataUtilTest(tf.test.TestCase):
+
+  def test_load_rgb_image(self):
+    image_path = os.path.join(FLAGS.test_srcdir, _TEST_DATA_DIR, 'test.jpg')
+    image_data = data_util.load_image(image_path)
+    self.assertEqual(image_data.shape, (5184, 3456, 3))
+
+
+if __name__ == '__main__':
+  tf.test.main()
diff --git a/mediapipe/model_maker/python/core/data/dataset.py b/mediapipe/model_maker/python/core/data/dataset.py
new file mode 100644
index 000000000..a92b05c0d
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/dataset.py
@@ -0,0 +1,164 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Common dataset for model training and evaluation."""
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import functools
+from typing import Callable, Optional, Tuple, TypeVar
+
+# Dependency imports
+import tensorflow as tf
+
+_DatasetT = TypeVar('_DatasetT', bound='Dataset')
+
+
+class Dataset(object):
+  """A generic dataset class for loading model training and evaluation dataset.
+
+  For each ML task, such as image classification, text classification etc., a
+  subclass can be derived from this class to provide task-specific data loading
+  utilities.
+  """
+
+  def __init__(self, tf_dataset: tf.data.Dataset, size: Optional[int] = None):
+    """Initializes Dataset class.
+
+    To build dataset from raw data, consider using the task specific utilities,
+    e.g. from_folder().
+
+    Args:
+      tf_dataset: A tf.data.Dataset object that contains a potentially large set
+        of elements, where each element is a pair of (input_data, target). The
+        `input_data` means the raw input data, like an image, a text etc., while
+        the `target` means the ground truth of the raw input data, e.g. the
+        classification label of the image etc.
+      size: The size of the dataset. tf.data.Dataset donesn't support a function
+        to get the length directly since it's lazy-loaded and may be infinite.
+    """
+    self._dataset = tf_dataset
+    self._size = size
+
+  @property
+  def size(self) -> Optional[int]:
+    """Returns the size of the dataset.
+
+    Note that this function may return None becuase the exact size of the
+    dataset isn't a necessary parameter to create an instance of this class,
+    and tf.data.Dataset donesn't support a function to get the length directly
+    since it's lazy-loaded and may be infinite.
+    In most cases, however, when an instance of this class is created by helper
+    functions like 'from_folder', the size of the dataset will be preprocessed,
+    and this function can return an int representing the size of the dataset.
+    """
+    return self._size
+
+  def gen_tf_dataset(self,
+                     batch_size: int = 1,
+                     is_training: bool = False,
+                     shuffle: bool = False,
+                     preprocess: Optional[Callable[..., bool]] = None,
+                     drop_remainder: bool = False) -> tf.data.Dataset:
+    """Generates a batched tf.data.Dataset for training/evaluation.
+
+    Args:
+      batch_size: An integer, the returned dataset will be batched by this size.
+      is_training: A boolean, when True, the returned dataset will be optionally
+        shuffled and repeated as an endless dataset.
+      shuffle: A boolean, when True, the returned dataset will be shuffled to
+        create randomness during model training.
+      preprocess: A function taking three arguments in order, feature, label and
+        boolean is_training.
+      drop_remainder: boolean, whether the finaly batch drops remainder.
+
+    Returns:
+      A TF dataset ready to be consumed by Keras model.
+    """
+    dataset = self._dataset
+
+    if preprocess:
+      preprocess = functools.partial(preprocess, is_training=is_training)
+      dataset = dataset.map(preprocess, num_parallel_calls=tf.data.AUTOTUNE)
+
+    if is_training:
+      if shuffle:
+        # Shuffle size should be bigger than the batch_size. Otherwise it's only
+        # shuffling within the batch, which equals to not having shuffle.
+        buffer_size = 3 * batch_size
+        # But since we are doing shuffle before repeat, it doesn't make sense to
+        # shuffle more than total available entries.
+        # TODO: Investigate if shuffling before / after repeat
+        # dataset can get a better performance?
+        # Shuffle after repeat will give a more randomized dataset and mix the
+        # epoch boundary: https://www.tensorflow.org/guide/data
+        if self._size:
+          buffer_size = min(self._size, buffer_size)
+        dataset = dataset.shuffle(buffer_size=buffer_size)
+
+    dataset = dataset.batch(batch_size, drop_remainder=drop_remainder)
+    dataset = dataset.prefetch(tf.data.AUTOTUNE)
+    # TODO: Consider converting dataset to distributed dataset
+    # here.
+    return dataset
+
+  def __len__(self):
+    """Returns the number of element of the dataset."""
+    if self._size is not None:
+      return self._size
+    else:
+      return len(self._dataset)
+
+  def split(self: _DatasetT, fraction: float) -> Tuple[_DatasetT, _DatasetT]:
+    """Splits dataset into two sub-datasets with the given fraction.
+
+    Primarily used for splitting the data set into training and testing sets.
+
+    Args:
+      fraction: A float value defines the fraction of the first returned
+        subdataset in the original data.
+
+    Returns:
+      The splitted two sub datasets.
+    """
+    return self._split(fraction)
+
+  def _split(self: _DatasetT, fraction: float,
+             *args) -> Tuple[_DatasetT, _DatasetT]:
+    """Implementation for `split` method and returns sub-class instances.
+
+    Child DataLoader classes, if requires additional constructor arguments,
+    should implement their own `split` method by calling `_split` with all
+    arguments to the constructor.
+
+    Args:
+      fraction: A float value defines the fraction of the first returned
+        subdataset in the original data.
+      *args: additional arguments passed to the sub-class constructor.
+
+    Returns:
+      The splitted two sub datasets.
+    """
+    assert (fraction > 0 and fraction < 1)
+
+    dataset = self._dataset
+
+    train_size = int(self._size * fraction)
+    trainset = self.__class__(dataset.take(train_size), train_size, *args)
+
+    test_size = self._size - train_size
+    testset = self.__class__(dataset.skip(train_size), test_size, *args)
+
+    return trainset, testset
diff --git a/mediapipe/model_maker/python/core/data/dataset_test.py b/mediapipe/model_maker/python/core/data/dataset_test.py
new file mode 100644
index 000000000..9adff127d
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/dataset_test.py
@@ -0,0 +1,78 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+# Dependency imports
+
+import numpy as np
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.data import dataset as ds
+from mediapipe.model_maker.python.core.utils import test_util
+
+
+class DatasetTest(tf.test.TestCase):
+
+  def test_split(self):
+    dataset = tf.data.Dataset.from_tensor_slices([[0, 1], [1, 1], [0, 0],
+                                                  [1, 0]])
+    data = ds.Dataset(dataset, 4)
+    train_data, test_data = data.split(0.5)
+
+    self.assertLen(train_data, 2)
+    self.assertIsInstance(train_data, ds.Dataset)
+    self.assertIsInstance(test_data, ds.Dataset)
+    for i, elem in enumerate(train_data.gen_tf_dataset()):
+      self.assertTrue((elem.numpy() == np.array([i, 1])).all())
+
+    self.assertLen(test_data, 2)
+    for i, elem in enumerate(test_data.gen_tf_dataset()):
+      self.assertTrue((elem.numpy() == np.array([i, 0])).all())
+
+  def test_len(self):
+    size = 4
+    dataset = tf.data.Dataset.from_tensor_slices([[0, 1], [1, 1], [0, 0],
+                                                  [1, 0]])
+    data = ds.Dataset(dataset, size)
+    self.assertLen(data, size)
+
+  def test_gen_tf_dataset(self):
+    input_dim = 8
+    data = test_util.create_dataset(
+        data_size=2, input_shape=[input_dim], num_classes=2)
+
+    dataset = data.gen_tf_dataset()
+    self.assertLen(dataset, 2)
+    for (feature, label) in dataset:
+      self.assertTrue((tf.shape(feature).numpy() == np.array([1, 8])).all())
+      self.assertTrue((tf.shape(label).numpy() == np.array([1])).all())
+
+    dataset2 = data.gen_tf_dataset(batch_size=2)
+    self.assertLen(dataset2, 1)
+    for (feature, label) in dataset2:
+      self.assertTrue((tf.shape(feature).numpy() == np.array([2, 8])).all())
+      self.assertTrue((tf.shape(label).numpy() == np.array([2])).all())
+
+    dataset3 = data.gen_tf_dataset(batch_size=2, is_training=True, shuffle=True)
+    self.assertEqual(dataset3.cardinality(), 1)
+    for (feature, label) in dataset3.take(10):
+      self.assertTrue((tf.shape(feature).numpy() == np.array([2, 8])).all())
+      self.assertTrue((tf.shape(label).numpy() == np.array([2])).all())
+
+
+if __name__ == '__main__':
+  tf.test.main()
diff --git a/mediapipe/model_maker/python/core/data/testdata/BUILD b/mediapipe/model_maker/python/core/data/testdata/BUILD
new file mode 100644
index 000000000..54e562d41
--- /dev/null
+++ b/mediapipe/model_maker/python/core/data/testdata/BUILD
@@ -0,0 +1,30 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+load(
+    "//mediapipe/framework/tool:mediapipe_files.bzl",
+    "mediapipe_files",
+)
+
+package(
+    default_visibility = ["//mediapipe/model_maker/python/core/data:__subpackages__"],
+    licenses = ["notice"],  # Apache 2.0
+)
+
+mediapipe_files(srcs = ["test.jpg"])
+
+filegroup(
+    name = "testdata",
+    srcs = ["test.jpg"],
+)
diff --git a/mediapipe/model_maker/python/core/utils/BUILD b/mediapipe/model_maker/python/core/utils/BUILD
new file mode 100644
index 000000000..e4b18b395
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/BUILD
@@ -0,0 +1,100 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+# Placeholder for internal Python strict library and test compatibility macro.
+
+licenses(["notice"])
+
+package(
+    default_visibility = ["//mediapipe:__subpackages__"],
+)
+
+py_library(
+    name = "test_util",
+    testonly = 1,
+    srcs = ["test_util.py"],
+    srcs_version = "PY3",
+    deps = [
+        ":model_util",
+        "//mediapipe/model_maker/python/core/data:dataset",
+    ],
+)
+
+py_library(
+    name = "image_preprocessing",
+    srcs = ["image_preprocessing.py"],
+    srcs_version = "PY3",
+)
+
+py_test(
+    name = "image_preprocessing_test",
+    srcs = ["image_preprocessing_test.py"],
+    python_version = "PY3",
+    srcs_version = "PY3",
+    deps = [":image_preprocessing"],
+)
+
+py_library(
+    name = "model_util",
+    srcs = ["model_util.py"],
+    srcs_version = "PY3",
+    deps = [
+        ":quantization",
+        "//mediapipe/model_maker/python/core/data:dataset",
+    ],
+)
+
+py_test(
+    name = "model_util_test",
+    srcs = ["model_util_test.py"],
+    python_version = "PY3",
+    srcs_version = "PY3",
+    deps = [
+        ":model_util",
+        ":quantization",
+        ":test_util",
+    ],
+)
+
+py_library(
+    name = "loss_functions",
+    srcs = ["loss_functions.py"],
+    srcs_version = "PY3",
+)
+
+py_test(
+    name = "loss_functions_test",
+    srcs = ["loss_functions_test.py"],
+    python_version = "PY3",
+    srcs_version = "PY3",
+    deps = [":loss_functions"],
+)
+
+py_library(
+    name = "quantization",
+    srcs = ["quantization.py"],
+    srcs_version = "PY3",
+    deps = ["//mediapipe/model_maker/python/core/data:dataset"],
+)
+
+py_test(
+    name = "quantization_test",
+    srcs = ["quantization_test.py"],
+    python_version = "PY3",
+    srcs_version = "PY3",
+    deps = [
+        ":quantization",
+        ":test_util",
+    ],
+)
diff --git a/mediapipe/model_maker/python/core/utils/__init__.py b/mediapipe/model_maker/python/core/utils/__init__.py
new file mode 100644
index 000000000..7ca2f9216
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/__init__.py
@@ -0,0 +1,13 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
diff --git a/mediapipe/model_maker/python/core/utils/image_preprocessing.py b/mediapipe/model_maker/python/core/utils/image_preprocessing.py
new file mode 100644
index 000000000..62b34fb27
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/image_preprocessing.py
@@ -0,0 +1,228 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+# ==============================================================================
+"""ImageNet preprocessing."""
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+# Dependency imports
+import tensorflow as tf
+
+IMAGE_SIZE = 224
+CROP_PADDING = 32
+
+
+class Preprocessor(object):
+  """Preprocessor for image classification."""
+
+  def __init__(self,
+               input_shape,
+               num_classes,
+               mean_rgb,
+               stddev_rgb,
+               use_augmentation=False):
+    self.input_shape = input_shape
+    self.num_classes = num_classes
+    self.mean_rgb = mean_rgb
+    self.stddev_rgb = stddev_rgb
+    self.use_augmentation = use_augmentation
+
+  def __call__(self, image, label, is_training=True):
+    if self.use_augmentation:
+      return self._preprocess_with_augmentation(image, label, is_training)
+    return self._preprocess_without_augmentation(image, label)
+
+  def _preprocess_with_augmentation(self, image, label, is_training):
+    """Image preprocessing method with data augmentation."""
+    image_size = self.input_shape[0]
+    if is_training:
+      image = preprocess_for_train(image, image_size)
+    else:
+      image = preprocess_for_eval(image, image_size)
+
+    image -= tf.constant(self.mean_rgb, shape=[1, 1, 3], dtype=image.dtype)
+    image /= tf.constant(self.stddev_rgb, shape=[1, 1, 3], dtype=image.dtype)
+
+    label = tf.one_hot(label, depth=self.num_classes)
+    return image, label
+
+  # TODO: Changes to preprocess to support batch input.
+  def _preprocess_without_augmentation(self, image, label):
+    """Image preprocessing method without data augmentation."""
+    image = tf.cast(image, tf.float32)
+
+    image -= tf.constant(self.mean_rgb, shape=[1, 1, 3], dtype=image.dtype)
+    image /= tf.constant(self.stddev_rgb, shape=[1, 1, 3], dtype=image.dtype)
+
+    image = tf.compat.v1.image.resize(image, self.input_shape)
+    label = tf.one_hot(label, depth=self.num_classes)
+    return image, label
+
+
+def _distorted_bounding_box_crop(image,
+                                 bbox,
+                                 min_object_covered=0.1,
+                                 aspect_ratio_range=(0.75, 1.33),
+                                 area_range=(0.05, 1.0),
+                                 max_attempts=100):
+  """Generates cropped_image using one of the bboxes randomly distorted.
+
+  See `tf.image.sample_distorted_bounding_box` for more documentation.
+
+  Args:
+    image: 4-D Tensor of shape [batch, height, width, channels] or 3-D Tensor of
+      shape [height, width, channels].
+    bbox: `Tensor` of bounding boxes arranged `[1, num_boxes, coords]` where
+      each coordinate is [0, 1) and the coordinates are arranged as `[ymin,
+      xmin, ymax, xmax]`. If num_boxes is 0 then use the whole image.
+    min_object_covered: An optional `float`. Defaults to `0.1`. The cropped area
+      of the image must contain at least this fraction of any bounding box
+      supplied.
+    aspect_ratio_range: An optional list of `float`s. The cropped area of the
+      image must have an aspect ratio = width / height within this range.
+    area_range: An optional list of `float`s. The cropped area of the image must
+      contain a fraction of the supplied image within in this range.
+    max_attempts: An optional `int`. Number of attempts at generating a cropped
+      region of the image of the specified constraints. After `max_attempts`
+      failures, return the entire image.
+
+  Returns:
+    A cropped image `Tensor`
+  """
+  with tf.name_scope('distorted_bounding_box_crop'):
+    shape = tf.shape(image)
+    sample_distorted_bounding_box = tf.image.sample_distorted_bounding_box(
+        shape,
+        bounding_boxes=bbox,
+        min_object_covered=min_object_covered,
+        aspect_ratio_range=aspect_ratio_range,
+        area_range=area_range,
+        max_attempts=max_attempts,
+        use_image_if_no_bounding_boxes=True)
+    bbox_begin, bbox_size, _ = sample_distorted_bounding_box
+
+    # Crop the image to the specified bounding box.
+    offset_y, offset_x, _ = tf.unstack(bbox_begin)
+    target_height, target_width, _ = tf.unstack(bbox_size)
+    image = tf.image.crop_to_bounding_box(image, offset_y, offset_x,
+                                          target_height, target_width)
+
+    return image
+
+
+def _at_least_x_are_equal(a, b, x):
+  """At least `x` of `a` and `b` `Tensors` are equal."""
+  match = tf.equal(a, b)
+  match = tf.cast(match, tf.int32)
+  return tf.greater_equal(tf.reduce_sum(match), x)
+
+
+def _resize_image(image, image_size, method=None):
+  if method is not None:
+    tf.compat.v1.logging.info('Use customized resize method {}'.format(method))
+    return tf.compat.v1.image.resize([image], [image_size, image_size],
+                                     method)[0]
+  tf.compat.v1.logging.info('Use default resize_bicubic.')
+  return tf.compat.v1.image.resize_bicubic([image], [image_size, image_size])[0]
+
+
+def _decode_and_random_crop(original_image, image_size, resize_method=None):
+  """Makes a random crop of image_size."""
+  bbox = tf.constant([0.0, 0.0, 1.0, 1.0], dtype=tf.float32, shape=[1, 1, 4])
+  image = _distorted_bounding_box_crop(
+      original_image,
+      bbox,
+      min_object_covered=0.1,
+      aspect_ratio_range=(3. / 4, 4. / 3.),
+      area_range=(0.08, 1.0),
+      max_attempts=10)
+  original_shape = tf.shape(original_image)
+  bad = _at_least_x_are_equal(original_shape, tf.shape(image), 3)
+
+  image = tf.cond(bad,
+                  lambda: _decode_and_center_crop(original_image, image_size),
+                  lambda: _resize_image(image, image_size, resize_method))
+
+  return image
+
+
+def _decode_and_center_crop(image, image_size, resize_method=None):
+  """Crops to center of image with padding then scales image_size."""
+  shape = tf.shape(image)
+  image_height = shape[0]
+  image_width = shape[1]
+
+  padded_center_crop_size = tf.cast(
+      ((image_size / (image_size + CROP_PADDING)) *
+       tf.cast(tf.minimum(image_height, image_width), tf.float32)), tf.int32)
+
+  offset_height = ((image_height - padded_center_crop_size) + 1) // 2
+  offset_width = ((image_width - padded_center_crop_size) + 1) // 2
+  image = tf.image.crop_to_bounding_box(image, offset_height, offset_width,
+                                        padded_center_crop_size,
+                                        padded_center_crop_size)
+  image = _resize_image(image, image_size, resize_method)
+  return image
+
+
+def _flip(image):
+  """Random horizontal image flip."""
+  image = tf.image.random_flip_left_right(image)
+  return image
+
+
+def preprocess_for_train(
+    image: tf.Tensor,
+    image_size: int = IMAGE_SIZE,
+    resize_method: str = tf.image.ResizeMethod.BILINEAR) -> tf.Tensor:
+  """Preprocesses the given image for evaluation.
+
+  Args:
+    image: 4-D Tensor of shape [batch, height, width, channels] or 3-D Tensor of
+      shape [height, width, channels].
+    image_size: image size.
+    resize_method: resize method. If none, use bicubic.
+
+  Returns:
+    A preprocessed image `Tensor`.
+  """
+  image = _decode_and_random_crop(image, image_size, resize_method)
+  image = _flip(image)
+  image = tf.reshape(image, [image_size, image_size, 3])
+
+  image = tf.image.convert_image_dtype(image, dtype=tf.float32)
+
+  return image
+
+
+def preprocess_for_eval(
+    image: tf.Tensor,
+    image_size: int = IMAGE_SIZE,
+    resize_method: str = tf.image.ResizeMethod.BILINEAR) -> tf.Tensor:
+  """Preprocesses the given image for evaluation.
+
+  Args:
+    image: 4-D Tensor of shape [batch, height, width, channels] or 3-D Tensor of
+      shape [height, width, channels].
+    image_size: image size.
+    resize_method: if None, use bicubic.
+
+  Returns:
+    A preprocessed image `Tensor`.
+  """
+  image = _decode_and_center_crop(image, image_size, resize_method)
+  image = tf.reshape(image, [image_size, image_size, 3])
+  image = tf.image.convert_image_dtype(image, dtype=tf.float32)
+  return image
diff --git a/mediapipe/model_maker/python/core/utils/image_preprocessing_test.py b/mediapipe/model_maker/python/core/utils/image_preprocessing_test.py
new file mode 100644
index 000000000..bc4b44569
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/image_preprocessing_test.py
@@ -0,0 +1,85 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the 'License');
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an 'AS IS' BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+# Dependency imports
+import numpy as np
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.utils import image_preprocessing
+
+
+def _get_preprocessed_image(preprocessor, is_training=False):
+  image_placeholder = tf.compat.v1.placeholder(tf.uint8, [24, 24, 3])
+  label_placeholder = tf.compat.v1.placeholder(tf.int32, [1])
+  image_tensor, _ = preprocessor(image_placeholder, label_placeholder,
+                                 is_training)
+
+  with tf.compat.v1.Session() as sess:
+    input_image = np.arange(24 * 24 * 3, dtype=np.uint8).reshape([24, 24, 3])
+    image = sess.run(
+        image_tensor,
+        feed_dict={
+            image_placeholder: input_image,
+            label_placeholder: [0]
+        })
+    return image
+
+
+class PreprocessorTest(tf.test.TestCase):
+
+  def test_preprocess_without_augmentation(self):
+    preprocessor = image_preprocessing.Preprocessor(input_shape=[2, 2],
+                                                    num_classes=2,
+                                                    mean_rgb=[0.0],
+                                                    stddev_rgb=[255.0],
+                                                    use_augmentation=False)
+    actual_image = np.array([[[0., 0.00392157, 0.00784314],
+                              [0.14117648, 0.14509805, 0.14901961]],
+                             [[0.37647063, 0.3803922, 0.38431376],
+                              [0.5176471, 0.52156866, 0.5254902]]])
+
+    image = _get_preprocessed_image(preprocessor)
+    self.assertTrue(np.allclose(image, actual_image, atol=1e-05))
+
+  def test_preprocess_with_augmentation(self):
+    image_preprocessing.CROP_PADDING = 1
+    preprocessor = image_preprocessing.Preprocessor(input_shape=[2, 2],
+                                                    num_classes=2,
+                                                    mean_rgb=[0.0],
+                                                    stddev_rgb=[255.0],
+                                                    use_augmentation=True)
+    # Tests validation image.
+    actual_eval_image = np.array([[[0.17254902, 0.1764706, 0.18039216],
+                                   [0.26666668, 0.27058825, 0.27450982]],
+                                  [[0.42352945, 0.427451, 0.43137258],
+                                   [0.5176471, 0.52156866, 0.5254902]]])
+
+    image = _get_preprocessed_image(preprocessor, is_training=False)
+    self.assertTrue(np.allclose(image, actual_eval_image, atol=1e-05))
+
+    # Tests training image.
+    image1 = _get_preprocessed_image(preprocessor, is_training=True)
+    image2 = _get_preprocessed_image(preprocessor, is_training=True)
+    self.assertFalse(np.allclose(image1, image2, atol=1e-05))
+    self.assertEqual(image1.shape, (2, 2, 3))
+    self.assertEqual(image2.shape, (2, 2, 3))
+
+
+if __name__ == '__main__':
+  tf.compat.v1.disable_eager_execution()
+  tf.test.main()
diff --git a/mediapipe/model_maker/python/core/utils/loss_functions.py b/mediapipe/model_maker/python/core/utils/loss_functions.py
new file mode 100644
index 000000000..17c738a14
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/loss_functions.py
@@ -0,0 +1,105 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Loss function utility library."""
+
+from typing import Optional, Sequence
+
+import tensorflow as tf
+
+
+class FocalLoss(tf.keras.losses.Loss):
+  """Implementation of focal loss (https://arxiv.org/pdf/1708.02002.pdf).
+
+  This class computes the focal loss between labels and prediction. Focal loss
+  is a weighted loss function that modulates the standard cross-entropy loss
+  based on how well the neural network performs on a specific example of a
+  class. The labels should be provided in a `one_hot` vector representation.
+  There should be `#classes` floating point values per prediction.
+  The loss is reduced across all samples using 'sum_over_batch_size' reduction
+  (see https://www.tensorflow.org/api_docs/python/tf/keras/losses/Reduction).
+
+  Example usage:
+  >>> y_true = [[0, 1, 0], [0, 0, 1]]
+  >>> y_pred = [[0.05, 0.95, 0], [0.1, 0.8, 0.1]]
+  >>> gamma = 2
+  >>> focal_loss = FocalLoss(gamma)
+  >>> focal_loss(y_true, y_pred).numpy()
+  0.9326
+
+  >>> # Calling with 'sample_weight'.
+  >>> focal_loss(y_true, y_pred, sample_weight=tf.constant([0.3, 0.7])).numpy()
+  0.6528
+
+  Usage with the `compile()` API:
+  ```python
+  model.compile(optimizer='sgd', loss=FocalLoss(gamma))
+  ```
+
+  """
+
+  def __init__(self, gamma, class_weight: Optional[Sequence[float]] = None):
+    """Constructor.
+
+    Args:
+      gamma: Focal loss gamma, as described in class docs.
+      class_weight: A weight to apply to the loss, one for each class. The
+        weight is applied for each input where the ground truth label matches.
+    """
+    super(tf.keras.losses.Loss, self).__init__()
+    # Used for clipping min/max values of probability values in y_pred to avoid
+    # NaNs and Infs in computation.
+    self._epsilon = 1e-7
+    # This is a tunable "focusing parameter"; should be >= 0.
+    # When gamma = 0, the loss returned is the standard categorical
+    # cross-entropy loss.
+    self._gamma = gamma
+    self._class_weight = class_weight
+    # tf.keras.losses.Loss class implementation requires a Reduction specified
+    # in self.reduction. To use this reduction, we should use tensorflow's
+    # compute_weighted_loss function however it is only compatible with v1 of
+    # Tensorflow: https://www.tensorflow.org/api_docs/python/tf/compat/v1/losses/compute_weighted_loss?hl=en.  pylint: disable=line-too-long
+    # So even though it is specified here, we don't use self.reduction in the
+    # loss function call.
+    self.reduction = tf.keras.losses.Reduction.SUM_OVER_BATCH_SIZE
+
+  def __call__(self,
+               y_true: tf.Tensor,
+               y_pred: tf.Tensor,
+               sample_weight: Optional[tf.Tensor] = None) -> tf.Tensor:
+    if self._class_weight:
+      class_weight = tf.convert_to_tensor(self._class_weight, dtype=tf.float32)
+      label = tf.argmax(y_true, axis=1)
+      loss_weight = tf.gather(class_weight, label)
+    else:
+      loss_weight = tf.ones(tf.shape(y_true)[0])
+    y_true = tf.cast(y_true, y_pred.dtype)
+    y_pred = tf.clip_by_value(y_pred, self._epsilon, 1 - self._epsilon)
+    batch_size = tf.cast(tf.shape(y_pred)[0], y_pred.dtype)
+    if sample_weight is None:
+      sample_weight = tf.constant(1.0)
+    weight_shape = sample_weight.shape
+    weight_rank = weight_shape.ndims
+    y_pred_rank = y_pred.shape.ndims
+    if y_pred_rank - weight_rank == 1:
+      sample_weight = tf.expand_dims(sample_weight, [-1])
+    elif weight_rank != 0:
+      raise ValueError(f'Unexpected sample_weights, should be either a scalar'
+                       f'or a vector of batch_size:{batch_size.numpy()}')
+    ce = -tf.math.log(y_pred)
+    modulating_factor = tf.math.pow(1 - y_pred, self._gamma)
+    losses = y_true * modulating_factor * ce * sample_weight
+    losses = losses * loss_weight[:, tf.newaxis]
+    # By default, this function uses "sum_over_batch_size" reduction for the
+    # loss per batch.
+    return tf.reduce_sum(losses) / batch_size
diff --git a/mediapipe/model_maker/python/core/utils/loss_functions_test.py b/mediapipe/model_maker/python/core/utils/loss_functions_test.py
new file mode 100644
index 000000000..716c329ef
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/loss_functions_test.py
@@ -0,0 +1,103 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import math
+from typing import Optional
+
+from absl.testing import parameterized
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.utils import loss_functions
+
+
+class LossFunctionsTest(tf.test.TestCase, parameterized.TestCase):
+
+  @parameterized.named_parameters(
+      dict(testcase_name='no_sample_weight', sample_weight=None),
+      dict(
+          testcase_name='with_sample_weight',
+          sample_weight=tf.constant([0.2, 0.2, 0.3, 0.1, 0.2])))
+  def test_focal_loss_gamma_0_is_cross_entropy(
+      self, sample_weight: Optional[tf.Tensor]):
+    y_true = tf.constant([[1, 0, 0], [0, 1, 0], [0, 0, 1], [1, 0, 0], [0, 1,
+                                                                       0]])
+    y_pred = tf.constant([[0.7, 0.1, 0.2], [0.6, 0.3, 0.1], [0.1, 0.5, 0.4],
+                          [0.8, 0.1, 0.1], [0.4, 0.5, 0.1]])
+
+    tf_cce = tf.keras.losses.CategoricalCrossentropy(
+        from_logits=False,
+        reduction=tf.keras.losses.Reduction.SUM_OVER_BATCH_SIZE)
+    focal_loss = loss_functions.FocalLoss(gamma=0)
+    self.assertAllClose(
+        tf_cce(y_true, y_pred, sample_weight=sample_weight),
+        focal_loss(y_true, y_pred, sample_weight=sample_weight), 1e-4)
+
+  def test_focal_loss_with_sample_weight(self):
+    y_true = tf.constant([[1, 0, 0], [0, 1, 0], [0, 0, 1], [1, 0, 0], [0, 1,
+                                                                       0]])
+    y_pred = tf.constant([[0.7, 0.1, 0.2], [0.6, 0.3, 0.1], [0.1, 0.5, 0.4],
+                          [0.8, 0.1, 0.1], [0.4, 0.5, 0.1]])
+
+    focal_loss = loss_functions.FocalLoss(gamma=0)
+
+    sample_weight = tf.constant([0.2, 0.2, 0.3, 0.1, 0.2])
+
+    self.assertGreater(
+        focal_loss(y_true=y_true, y_pred=y_pred),
+        focal_loss(y_true=y_true, y_pred=y_pred, sample_weight=sample_weight))
+
+  @parameterized.named_parameters(
+      dict(testcase_name='gt_0.1', y_pred=tf.constant([0.1, 0.9])),
+      dict(testcase_name='gt_0.3', y_pred=tf.constant([0.3, 0.7])),
+      dict(testcase_name='gt_0.5', y_pred=tf.constant([0.5, 0.5])),
+      dict(testcase_name='gt_0.7', y_pred=tf.constant([0.7, 0.3])),
+      dict(testcase_name='gt_0.9', y_pred=tf.constant([0.9, 0.1])),
+  )
+  def test_focal_loss_decreases_with_increasing_gamma(self, y_pred: tf.Tensor):
+    y_true = tf.constant([[1, 0]])
+
+    focal_loss_gamma_0 = loss_functions.FocalLoss(gamma=0)
+    loss_gamma_0 = focal_loss_gamma_0(y_true, y_pred)
+    focal_loss_gamma_0p5 = loss_functions.FocalLoss(gamma=0.5)
+    loss_gamma_0p5 = focal_loss_gamma_0p5(y_true, y_pred)
+    focal_loss_gamma_1 = loss_functions.FocalLoss(gamma=1)
+    loss_gamma_1 = focal_loss_gamma_1(y_true, y_pred)
+    focal_loss_gamma_2 = loss_functions.FocalLoss(gamma=2)
+    loss_gamma_2 = focal_loss_gamma_2(y_true, y_pred)
+    focal_loss_gamma_5 = loss_functions.FocalLoss(gamma=5)
+    loss_gamma_5 = focal_loss_gamma_5(y_true, y_pred)
+
+    self.assertGreater(loss_gamma_0, loss_gamma_0p5)
+    self.assertGreater(loss_gamma_0p5, loss_gamma_1)
+    self.assertGreater(loss_gamma_1, loss_gamma_2)
+    self.assertGreater(loss_gamma_2, loss_gamma_5)
+
+  @parameterized.named_parameters(
+      dict(testcase_name='index_0', true_class=0),
+      dict(testcase_name='index_1', true_class=1),
+      dict(testcase_name='index_2', true_class=2),
+  )
+  def test_focal_loss_class_weight_is_applied(self, true_class: int):
+    class_weight = [1.0, 3.0, 10.0]
+    y_pred = tf.constant([[1.0, 1.0, 1.0]]) / 3.0
+    y_true = tf.one_hot(true_class, depth=3)[tf.newaxis, :]
+    expected_loss = -math.log(1.0 / 3.0) * class_weight[true_class]
+
+    loss_fn = loss_functions.FocalLoss(gamma=0, class_weight=class_weight)
+    loss = loss_fn(y_true, y_pred)
+    self.assertNear(loss, expected_loss, 1e-4)
+
+
+if __name__ == '__main__':
+  tf.test.main()
diff --git a/mediapipe/model_maker/python/core/utils/model_util.py b/mediapipe/model_maker/python/core/utils/model_util.py
new file mode 100644
index 000000000..4914fea57
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/model_util.py
@@ -0,0 +1,241 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Utilities for keras models."""
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+import os
+import tempfile
+from typing import Any, Callable, Dict, List, Optional, Text, Tuple, Union
+
+# Dependency imports
+
+import numpy as np
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.data import dataset
+from mediapipe.model_maker.python.core.utils import quantization
+
+DEFAULT_SCALE, DEFAULT_ZERO_POINT = 0, 0
+ESTIMITED_STEPS_PER_EPOCH = 1000
+
+
+def get_steps_per_epoch(steps_per_epoch: Optional[int] = None,
+                        batch_size: Optional[int] = None,
+                        train_data: Optional[dataset.Dataset] = None) -> int:
+  """Gets the estimated training steps per epoch.
+
+  1. If `steps_per_epoch` is set, returns `steps_per_epoch` directly.
+  2. Else if we can get the length of training data successfully, returns
+     `train_data_length // batch_size`.
+
+  Args:
+    steps_per_epoch: int, training steps per epoch.
+    batch_size: int, batch size.
+    train_data: training data.
+
+  Returns:
+    Estimated training steps per epoch.
+
+  Raises:
+    ValueError: if both steps_per_epoch and train_data are not set.
+  """
+  if steps_per_epoch is not None:
+    # steps_per_epoch is set by users manually.
+    return steps_per_epoch
+  else:
+    if train_data is None:
+      raise ValueError('Input train_data cannot be None.')
+    # Gets the steps by the length of the training data.
+    return len(train_data) // batch_size
+
+
+def export_tflite(
+    model: tf.keras.Model,
+    tflite_filepath: str,
+    quantization_config: Optional[quantization.QuantizationConfig] = None,
+    supported_ops: Tuple[tf.lite.OpsSet,
+                         ...] = (tf.lite.OpsSet.TFLITE_BUILTINS,)):
+  """Converts the model to tflite format and saves it.
+
+  Args:
+    model: model to be converted to tflite.
+    tflite_filepath: File path to save tflite model.
+    quantization_config: Configuration for post-training quantization.
+    supported_ops: A list of supported ops in the converted TFLite file.
+  """
+  if tflite_filepath is None:
+    raise ValueError(
+        "TFLite filepath couldn't be None when exporting to tflite.")
+
+  with tempfile.TemporaryDirectory() as temp_dir:
+    save_path = os.path.join(temp_dir, 'saved_model')
+    model.save(save_path, include_optimizer=False, save_format='tf')
+    converter = tf.lite.TFLiteConverter.from_saved_model(save_path)
+
+    if quantization_config:
+      converter = quantization_config.set_converter_with_quantization(converter)
+
+    converter.target_spec.supported_ops = supported_ops
+    tflite_model = converter.convert()
+
+  with tf.io.gfile.GFile(tflite_filepath, 'wb') as f:
+    f.write(tflite_model)
+
+
+class WarmUp(tf.keras.optimizers.schedules.LearningRateSchedule):
+  """Applies a warmup schedule on a given learning rate decay schedule."""
+
+  def __init__(self,
+               initial_learning_rate: float,
+               decay_schedule_fn: Callable[[Any], Any],
+               warmup_steps: int,
+               name: Optional[str] = None):
+    """Initializes a new instance of the `WarmUp` class.
+
+    Args:
+      initial_learning_rate: learning rate after the warmup.
+      decay_schedule_fn: A function maps step to learning rate. Will be applied
+        for values of step larger than 'warmup_steps'.
+      warmup_steps: Number of steps to do warmup for.
+      name: TF namescope under which to perform the learning rate calculation.
+    """
+    super(WarmUp, self).__init__()
+    self.initial_learning_rate = initial_learning_rate
+    self.warmup_steps = warmup_steps
+    self.decay_schedule_fn = decay_schedule_fn
+    self.name = name
+
+  def __call__(self, step: Union[int, tf.Tensor]) -> tf.Tensor:
+    with tf.name_scope(self.name or 'WarmUp') as name:
+      # Implements linear warmup. i.e., if global_step < warmup_steps, the
+      # learning rate will be `global_step/num_warmup_steps * init_lr`.
+      global_step_float = tf.cast(step, tf.float32)
+      warmup_steps_float = tf.cast(self.warmup_steps, tf.float32)
+      warmup_percent_done = global_step_float / warmup_steps_float
+      warmup_learning_rate = self.initial_learning_rate * warmup_percent_done
+      return tf.cond(
+          global_step_float < warmup_steps_float,
+          lambda: warmup_learning_rate,
+          lambda: self.decay_schedule_fn(step),
+          name=name)
+
+  def get_config(self) -> Dict[Text, Any]:
+    return {
+        'initial_learning_rate': self.initial_learning_rate,
+        'decay_schedule_fn': self.decay_schedule_fn,
+        'warmup_steps': self.warmup_steps,
+        'name': self.name
+    }
+
+
+class LiteRunner(object):
+  """A runner to do inference with the TFLite model."""
+
+  def __init__(self, tflite_filepath: str):
+    """Initializes Lite runner with tflite model file.
+
+    Args:
+      tflite_filepath: File path to the TFLite model.
+    """
+    with tf.io.gfile.GFile(tflite_filepath, 'rb') as f:
+      tflite_model = f.read()
+    self.interpreter = tf.lite.Interpreter(model_content=tflite_model)
+    self.interpreter.allocate_tensors()
+    self.input_details = self.interpreter.get_input_details()
+    self.output_details = self.interpreter.get_output_details()
+
+  def run(
+      self, input_tensors: Union[List[tf.Tensor], Dict[str, tf.Tensor]]
+  ) -> Union[List[tf.Tensor], tf.Tensor]:
+    """Runs inference with the TFLite model.
+
+    Args:
+      input_tensors: List / Dict of the input tensors of the TFLite model. The
+        order should be the same as the keras model if it's a list. It also
+        accepts tensor directly if the model has only 1 input.
+
+    Returns:
+      List of the output tensors for multi-output models, otherwise just
+        the output tensor. The order should be the same as the keras model.
+    """
+
+    if not isinstance(input_tensors, list) and not isinstance(
+        input_tensors, dict):
+      input_tensors = [input_tensors]
+
+    interpreter = self.interpreter
+
+    # Reshape inputs
+    for i, input_detail in enumerate(self.input_details):
+      input_tensor = _get_input_tensor(
+          input_tensors=input_tensors,
+          input_details=self.input_details,
+          index=i)
+      interpreter.resize_tensor_input(
+          input_index=input_detail['index'], tensor_size=input_tensor.shape)
+    interpreter.allocate_tensors()
+
+    # Feed input to the interpreter
+    for i, input_detail in enumerate(self.input_details):
+      input_tensor = _get_input_tensor(
+          input_tensors=input_tensors,
+          input_details=self.input_details,
+          index=i)
+      if input_detail['quantization'] != (DEFAULT_SCALE, DEFAULT_ZERO_POINT):
+        # Quantize the input
+        scale, zero_point = input_detail['quantization']
+        input_tensor = input_tensor / scale + zero_point
+        input_tensor = np.array(input_tensor, dtype=input_detail['dtype'])
+      interpreter.set_tensor(input_detail['index'], input_tensor)
+
+    interpreter.invoke()
+
+    output_tensors = []
+    for output_detail in self.output_details:
+      output_tensor = interpreter.get_tensor(output_detail['index'])
+      if output_detail['quantization'] != (DEFAULT_SCALE, DEFAULT_ZERO_POINT):
+        # Dequantize the output
+        scale, zero_point = output_detail['quantization']
+        output_tensor = output_tensor.astype(np.float32)
+        output_tensor = (output_tensor - zero_point) * scale
+      output_tensors.append(output_tensor)
+
+    if len(output_tensors) == 1:
+      return output_tensors[0]
+    return output_tensors
+
+
+def get_lite_runner(tflite_filepath: str) -> 'LiteRunner':
+  """Returns a `LiteRunner` from file path to TFLite model."""
+  lite_runner = LiteRunner(tflite_filepath)
+  return lite_runner
+
+
+def _get_input_tensor(input_tensors: Union[List[tf.Tensor], Dict[str,
+                                                                 tf.Tensor]],
+                      input_details: Dict[str, Any], index: int) -> tf.Tensor:
+  """Returns input tensor in `input_tensors` that maps `input_detail[i]`."""
+  if isinstance(input_tensors, dict):
+    # Gets the mapped input tensor.
+    input_detail = input_details
+    for input_tensor_name, input_tensor in input_tensors.items():
+      if input_tensor_name in input_detail['name']:
+        return input_tensor
+    raise ValueError('Input tensors don\'t contains a tensor that mapped the '
+                     'input detail %s' % str(input_detail))
+  else:
+    return input_tensors[index]
diff --git a/mediapipe/model_maker/python/core/utils/model_util_test.py b/mediapipe/model_maker/python/core/utils/model_util_test.py
new file mode 100644
index 000000000..9c3908841
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/model_util_test.py
@@ -0,0 +1,137 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import os
+
+from absl.testing import parameterized
+import numpy as np
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.utils import model_util
+from mediapipe.model_maker.python.core.utils import quantization
+from mediapipe.model_maker.python.core.utils import test_util
+
+
+class ModelUtilTest(tf.test.TestCase, parameterized.TestCase):
+
+  @parameterized.named_parameters(
+      dict(
+          testcase_name='input_only_steps_per_epoch',
+          steps_per_epoch=1000,
+          batch_size=None,
+          train_data=None,
+          expected_steps_per_epoch=1000),
+      dict(
+          testcase_name='input_steps_per_epoch_and_batch_size',
+          steps_per_epoch=1000,
+          batch_size=32,
+          train_data=None,
+          expected_steps_per_epoch=1000),
+      dict(
+          testcase_name='input_steps_per_epoch_batch_size_and_train_data',
+          steps_per_epoch=1000,
+          batch_size=32,
+          train_data=tf.data.Dataset.from_tensor_slices([[0, 1], [1, 1], [0, 0],
+                                                         [1, 0]]),
+          expected_steps_per_epoch=1000),
+      dict(
+          testcase_name='input_batch_size_and_train_data',
+          steps_per_epoch=None,
+          batch_size=2,
+          train_data=tf.data.Dataset.from_tensor_slices([[0, 1], [1, 1], [0, 0],
+                                                         [1, 0]]),
+          expected_steps_per_epoch=2))
+  def test_get_steps_per_epoch(self, steps_per_epoch, batch_size, train_data,
+                               expected_steps_per_epoch):
+    estimated_steps_per_epoch = model_util.get_steps_per_epoch(
+        steps_per_epoch=steps_per_epoch,
+        batch_size=batch_size,
+        train_data=train_data)
+    self.assertEqual(estimated_steps_per_epoch, expected_steps_per_epoch)
+
+  def test_get_steps_per_epoch_raise_value_error(self):
+    with self.assertRaises(ValueError):
+      model_util.get_steps_per_epoch(
+          steps_per_epoch=None, batch_size=16, train_data=None)
+
+  def test_warmup(self):
+    init_lr = 0.1
+    warmup_steps = 1000
+    num_decay_steps = 100
+    learning_rate_fn = tf.keras.experimental.CosineDecay(
+        initial_learning_rate=init_lr, decay_steps=num_decay_steps)
+    warmup_object = model_util.WarmUp(
+        initial_learning_rate=init_lr,
+        decay_schedule_fn=learning_rate_fn,
+        warmup_steps=1000,
+        name='test')
+    self.assertEqual(
+        warmup_object.get_config(), {
+            'initial_learning_rate': init_lr,
+            'decay_schedule_fn': learning_rate_fn,
+            'warmup_steps': warmup_steps,
+            'name': 'test'
+        })
+
+  def test_export_tflite(self):
+    input_dim = 4
+    model = test_util.build_model(input_shape=[input_dim], num_classes=2)
+    tflite_file = os.path.join(self.get_temp_dir(), 'model.tflite')
+    model_util.export_tflite(model, tflite_file)
+    self._test_tflite(model, tflite_file, input_dim)
+
+  @parameterized.named_parameters(
+      dict(
+          testcase_name='dynamic_quantize',
+          config=quantization.QuantizationConfig.for_dynamic(),
+          model_size=1288),
+      dict(
+          testcase_name='int8_quantize',
+          config=quantization.QuantizationConfig.for_int8(
+              representative_data=test_util.create_dataset(
+                  data_size=10, input_shape=[16], num_classes=3)),
+          model_size=1832),
+      dict(
+          testcase_name='float16_quantize',
+          config=quantization.QuantizationConfig.for_float16(),
+          model_size=1468))
+  def test_export_tflite_quantized(self, config, model_size):
+    input_dim = 16
+    num_classes = 2
+    max_input_value = 5
+    model = test_util.build_model([input_dim], num_classes)
+    tflite_file = os.path.join(self.get_temp_dir(), 'model_quantized.tflite')
+
+    model_util.export_tflite(model, tflite_file, config)
+    self._test_tflite(
+        model, tflite_file, input_dim, max_input_value, atol=1e-00)
+    self.assertNear(os.path.getsize(tflite_file), model_size, 300)
+
+  def _test_tflite(self,
+                   keras_model: tf.keras.Model,
+                   tflite_model_file: str,
+                   input_dim: int,
+                   max_input_value: int = 1000,
+                   atol: float = 1e-04):
+    np.random.seed(0)
+    random_input = np.random.uniform(
+        low=0, high=max_input_value, size=(1, input_dim)).astype(np.float32)
+
+    self.assertTrue(
+        test_util.is_same_output(
+            tflite_model_file, keras_model, random_input, atol=atol))
+
+
+if __name__ == '__main__':
+  tf.test.main()
diff --git a/mediapipe/model_maker/python/core/utils/quantization.py b/mediapipe/model_maker/python/core/utils/quantization.py
new file mode 100644
index 000000000..a1a38cc64
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/quantization.py
@@ -0,0 +1,213 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Libraries for post-training quantization."""
+
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+from typing import Any, Callable, List, Optional, Union
+
+# Dependency imports
+
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.data import dataset as ds
+
+DEFAULT_QUANTIZATION_STEPS = 500
+
+
+def _get_representative_dataset_generator(dataset: tf.data.Dataset,
+                                          num_steps: int) -> Callable[[], Any]:
+  """Gets a representative dataset generator for post-training quantization.
+
+  The generator is to provide a small dataset to calibrate or estimate the
+  range, i.e, (min, max) of all floating-point arrays in the model for
+  quantization. Usually, this is a small subset of a few hundred samples
+  randomly chosen, in no particular order, from the training or evaluation
+  dataset. See tf.lite.RepresentativeDataset for more details.
+
+  Args:
+    dataset: Input dataset for extracting representative sub dataset.
+    num_steps: The number of quantization steps which also reflects the size of
+      the representative dataset.
+
+  Returns:
+    A representative dataset generator.
+  """
+
+  def representative_dataset_gen():
+    """Generates representative dataset for quantization."""
+    for data, _ in dataset.take(num_steps):
+      yield [data]
+
+  return representative_dataset_gen
+
+
+class QuantizationConfig(object):
+  """Configuration for post-training quantization.
+
+  Refer to
+  https://www.tensorflow.org/lite/performance/post_training_quantization
+  for different post-training quantization options.
+  """
+
+  def __init__(
+      self,
+      optimizations: Optional[Union[tf.lite.Optimize,
+                                    List[tf.lite.Optimize]]] = None,
+      representative_data: Optional[ds.Dataset] = None,
+      quantization_steps: Optional[int] = None,
+      inference_input_type: Optional[tf.dtypes.DType] = None,
+      inference_output_type: Optional[tf.dtypes.DType] = None,
+      supported_ops: Optional[Union[tf.lite.OpsSet,
+                                    List[tf.lite.OpsSet]]] = None,
+      supported_types: Optional[Union[tf.dtypes.DType,
+                                      List[tf.dtypes.DType]]] = None,
+      experimental_new_quantizer: bool = False,
+  ):
+    """Constructs QuantizationConfig.
+
+    Args:
+      optimizations: A list of optimizations to apply when converting the model.
+        If not set, use `[Optimize.DEFAULT]` by default.
+      representative_data: A representative ds.Dataset for post-training
+        quantization.
+      quantization_steps: Number of post-training quantization calibration steps
+        to run (default to DEFAULT_QUANTIZATION_STEPS).
+      inference_input_type: Target data type of real-number input arrays. Allows
+        for a different type for input arrays. Defaults to None. If set, must be
+        be `{tf.float32, tf.uint8, tf.int8}`.
+      inference_output_type: Target data type of real-number output arrays.
+        Allows for a different type for output arrays. Defaults to None. If set,
+        must be `{tf.float32, tf.uint8, tf.int8}`.
+      supported_ops: Set of OpsSet options supported by the device. Used to Set
+        converter.target_spec.supported_ops.
+      supported_types: List of types for constant values on the target device.
+        Supported values are types exported by lite.constants. Frequently, an
+        optimization choice is driven by the most compact (i.e. smallest) type
+        in this list (default [constants.FLOAT]).
+      experimental_new_quantizer: Whether to enable experimental new quantizer.
+
+    Raises:
+      ValueError: if inference_input_type or inference_output_type are set but
+      not in {tf.float32, tf.uint8, tf.int8}.
+    """
+    if inference_input_type is not None and inference_input_type not in {
+        tf.float32, tf.uint8, tf.int8
+    }:
+      raise ValueError('Unsupported inference_input_type %s' %
+                       inference_input_type)
+    if inference_output_type is not None and inference_output_type not in {
+        tf.float32, tf.uint8, tf.int8
+    }:
+      raise ValueError('Unsupported inference_output_type %s' %
+                       inference_output_type)
+
+    if optimizations is None:
+      optimizations = [tf.lite.Optimize.DEFAULT]
+    if not isinstance(optimizations, list):
+      optimizations = [optimizations]
+    self.optimizations = optimizations
+
+    self.representative_data = representative_data
+    if self.representative_data is not None and quantization_steps is None:
+      quantization_steps = DEFAULT_QUANTIZATION_STEPS
+    self.quantization_steps = quantization_steps
+
+    self.inference_input_type = inference_input_type
+    self.inference_output_type = inference_output_type
+
+    if supported_ops is not None and not isinstance(supported_ops, list):
+      supported_ops = [supported_ops]
+    self.supported_ops = supported_ops
+
+    if supported_types is not None and not isinstance(supported_types, list):
+      supported_types = [supported_types]
+    self.supported_types = supported_types
+
+    self.experimental_new_quantizer = experimental_new_quantizer
+
+  @classmethod
+  def for_dynamic(cls) -> 'QuantizationConfig':
+    """Creates configuration for dynamic range quantization."""
+    return QuantizationConfig()
+
+  @classmethod
+  def for_int8(
+      cls,
+      representative_data: ds.Dataset,
+      quantization_steps: int = DEFAULT_QUANTIZATION_STEPS,
+      inference_input_type: tf.dtypes.DType = tf.uint8,
+      inference_output_type: tf.dtypes.DType = tf.uint8,
+      supported_ops: tf.lite.OpsSet = tf.lite.OpsSet.TFLITE_BUILTINS_INT8
+  ) -> 'QuantizationConfig':
+    """Creates configuration for full integer quantization.
+
+    Args:
+      representative_data: Representative data used for post-training
+        quantization.
+      quantization_steps: Number of post-training quantization calibration steps
+        to run.
+      inference_input_type: Target data type of real-number input arrays.
+      inference_output_type: Target data type of real-number output arrays.
+      supported_ops: Set of `tf.lite.OpsSet` options, where each option
+        represents a set of operators supported by the target device.
+
+    Returns:
+      QuantizationConfig.
+    """
+    return QuantizationConfig(
+        representative_data=representative_data,
+        quantization_steps=quantization_steps,
+        inference_input_type=inference_input_type,
+        inference_output_type=inference_output_type,
+        supported_ops=supported_ops)
+
+  @classmethod
+  def for_float16(cls) -> 'QuantizationConfig':
+    """Creates configuration for float16 quantization."""
+    return QuantizationConfig(supported_types=[tf.float16])
+
+  def set_converter_with_quantization(self, converter: tf.lite.TFLiteConverter,
+                                      **kwargs: Any) -> tf.lite.TFLiteConverter:
+    """Sets input TFLite converter with quantization configurations.
+
+    Args:
+      converter: input tf.lite.TFLiteConverter.
+      **kwargs: arguments used by ds.Dataset.gen_tf_dataset.
+
+    Returns:
+      tf.lite.TFLiteConverter with quantization configurations.
+    """
+    converter.optimizations = self.optimizations
+
+    if self.representative_data is not None:
+      tf_ds = self.representative_data.gen_tf_dataset(
+          batch_size=1, is_training=False, **kwargs)
+      converter.representative_dataset = tf.lite.RepresentativeDataset(
+          _get_representative_dataset_generator(tf_ds, self.quantization_steps))
+
+    if self.inference_input_type:
+      converter.inference_input_type = self.inference_input_type
+    if self.inference_output_type:
+      converter.inference_output_type = self.inference_output_type
+    if self.supported_ops:
+      converter.target_spec.supported_ops = self.supported_ops
+    if self.supported_types:
+      converter.target_spec.supported_types = self.supported_types
+
+    if self.experimental_new_quantizer is not None:
+      converter.experimental_new_quantizer = self.experimental_new_quantizer
+    return converter
diff --git a/mediapipe/model_maker/python/core/utils/quantization_test.py b/mediapipe/model_maker/python/core/utils/quantization_test.py
new file mode 100644
index 000000000..9d27d34ac
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/quantization_test.py
@@ -0,0 +1,108 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from absl.testing import parameterized
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.utils import quantization
+from mediapipe.model_maker.python.core.utils import test_util
+
+
+class QuantizationTest(tf.test.TestCase, parameterized.TestCase):
+
+  def test_create_dynamic_quantization_config(self):
+    config = quantization.QuantizationConfig.for_dynamic()
+    self.assertEqual(config.optimizations, [tf.lite.Optimize.DEFAULT])
+    self.assertIsNone(config.representative_data)
+    self.assertIsNone(config.inference_input_type)
+    self.assertIsNone(config.inference_output_type)
+    self.assertIsNone(config.supported_ops)
+    self.assertIsNone(config.supported_types)
+    self.assertFalse(config.experimental_new_quantizer)
+
+  def test_create_int8_quantization_config(self):
+    representative_data = test_util.create_dataset(
+        data_size=10, input_shape=[4], num_classes=3)
+    config = quantization.QuantizationConfig.for_int8(
+        representative_data=representative_data)
+    self.assertEqual(config.optimizations, [tf.lite.Optimize.DEFAULT])
+    self.assertEqual(config.inference_input_type, tf.uint8)
+    self.assertEqual(config.inference_output_type, tf.uint8)
+    self.assertEqual(config.supported_ops,
+                     [tf.lite.OpsSet.TFLITE_BUILTINS_INT8])
+    self.assertFalse(config.experimental_new_quantizer)
+
+  def test_set_converter_with_quantization_from_int8_config(self):
+    representative_data = test_util.create_dataset(
+        data_size=10, input_shape=[4], num_classes=3)
+    config = quantization.QuantizationConfig.for_int8(
+        representative_data=representative_data)
+    model = test_util.build_model(input_shape=[4], num_classes=3)
+    saved_model_dir = self.get_temp_dir()
+    model.save(saved_model_dir)
+    converter = tf.lite.TFLiteConverter.from_saved_model(saved_model_dir)
+    converter = config.set_converter_with_quantization(converter=converter)
+    self.assertEqual(config.optimizations, [tf.lite.Optimize.DEFAULT])
+    self.assertEqual(config.inference_input_type, tf.uint8)
+    self.assertEqual(config.inference_output_type, tf.uint8)
+    self.assertEqual(config.supported_ops,
+                     [tf.lite.OpsSet.TFLITE_BUILTINS_INT8])
+    tflite_model = converter.convert()
+    interpreter = tf.lite.Interpreter(model_content=tflite_model)
+    self.assertEqual(interpreter.get_input_details()[0]['dtype'], tf.uint8)
+    self.assertEqual(interpreter.get_output_details()[0]['dtype'], tf.uint8)
+
+  def test_create_float16_quantization_config(self):
+    config = quantization.QuantizationConfig.for_float16()
+    self.assertEqual(config.optimizations, [tf.lite.Optimize.DEFAULT])
+    self.assertIsNone(config.representative_data)
+    self.assertIsNone(config.inference_input_type)
+    self.assertIsNone(config.inference_output_type)
+    self.assertIsNone(config.supported_ops)
+    self.assertEqual(config.supported_types, [tf.float16])
+    self.assertFalse(config.experimental_new_quantizer)
+
+  def test_set_converter_with_quantization_from_float16_config(self):
+    config = quantization.QuantizationConfig.for_float16()
+    model = test_util.build_model(input_shape=[4], num_classes=3)
+    saved_model_dir = self.get_temp_dir()
+    model.save(saved_model_dir)
+    converter = tf.lite.TFLiteConverter.from_saved_model(saved_model_dir)
+    converter = config.set_converter_with_quantization(converter=converter)
+    self.assertEqual(config.supported_types, [tf.float16])
+    tflite_model = converter.convert()
+    interpreter = tf.lite.Interpreter(model_content=tflite_model)
+    # The input and output are expected to be set to float32 by default.
+    self.assertEqual(interpreter.get_input_details()[0]['dtype'], tf.float32)
+    self.assertEqual(interpreter.get_output_details()[0]['dtype'], tf.float32)
+
+  @parameterized.named_parameters(
+      dict(
+          testcase_name='invalid_inference_input_type',
+          inference_input_type=tf.uint8,
+          inference_output_type=tf.int64),
+      dict(
+          testcase_name='invalid_inference_output_type',
+          inference_input_type=tf.int64,
+          inference_output_type=tf.float32))
+  def test_create_quantization_config_failure(self, inference_input_type,
+                                              inference_output_type):
+    with self.assertRaises(ValueError):
+      _ = quantization.QuantizationConfig(
+          inference_input_type=inference_input_type,
+          inference_output_type=inference_output_type)
+
+
+if __name__ == '__main__':
+  tf.test.main()
diff --git a/mediapipe/model_maker/python/core/utils/test_util.py b/mediapipe/model_maker/python/core/utils/test_util.py
new file mode 100644
index 000000000..eb2952dd3
--- /dev/null
+++ b/mediapipe/model_maker/python/core/utils/test_util.py
@@ -0,0 +1,76 @@
+# Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+"""Test utilities for model maker."""
+from __future__ import absolute_import
+from __future__ import division
+from __future__ import print_function
+
+from typing import List, Union
+
+# Dependency imports
+
+import numpy as np
+import tensorflow as tf
+
+from mediapipe.model_maker.python.core.data import dataset as ds
+from mediapipe.model_maker.python.core.utils import model_util
+
+
+def create_dataset(data_size: int,
+                   input_shape: List[int],
+                   num_classes: int,
+                   max_input_value: int = 1000) -> ds.Dataset:
+  """Creates and returns a simple `Dataset` object for test."""
+  features = tf.random.uniform(
+      shape=[data_size] + input_shape,
+      minval=0,
+      maxval=max_input_value,
+      dtype=tf.float32)
+
+  labels = tf.random.uniform(
+      shape=[data_size], minval=0, maxval=num_classes, dtype=tf.int32)
+
+  tf_dataset = tf.data.Dataset.from_tensor_slices((features, labels))
+  dataset = ds.Dataset(tf_dataset, data_size)
+  return dataset
+
+
+def build_model(input_shape: List[int], num_classes: int) -> tf.keras.Model:
+  """Builds a simple Keras model for test."""
+  inputs = tf.keras.layers.Input(shape=input_shape)
+  if len(input_shape) == 3:  # Image inputs.
+    outputs = tf.keras.layers.GlobalAveragePooling2D()(inputs)
+    outputs = tf.keras.layers.Dense(num_classes, activation="softmax")(outputs)
+  elif len(input_shape) == 1:  # Text inputs.
+    outputs = tf.keras.layers.Dense(num_classes, activation="softmax")(inputs)
+  else:
+    raise ValueError("Model inputs should be 2D tensor or 4D tensor.")
+
+  model = tf.keras.Model(inputs=inputs, outputs=outputs)
+  return model
+
+
+def is_same_output(tflite_file: str,
+                   keras_model: tf.keras.Model,
+                   input_tensors: Union[List[tf.Tensor], tf.Tensor],
+                   atol: float = 1e-04) -> bool:
+  """Returns if the output of TFLite model and keras model are identical."""
+  # Gets output from lite model.
+  lite_runner = model_util.get_lite_runner(tflite_file)
+  lite_output = lite_runner.run(input_tensors)
+
+  # Gets output from keras model.
+  keras_output = keras_model.predict_on_batch(input_tensors)
+
+  return np.allclose(lite_output, keras_output, atol=atol)
diff --git a/mediapipe/model_maker/requirements.txt b/mediapipe/model_maker/requirements.txt
new file mode 100644
index 000000000..5e3832b09
--- /dev/null
+++ b/mediapipe/model_maker/requirements.txt
@@ -0,0 +1,4 @@
+absl-py
+numpy
+opencv-contrib-python
+tensorflow
diff --git a/third_party/external_files.bzl b/third_party/external_files.bzl
index cd291fc1e..24ceba639 100644
--- a/third_party/external_files.bzl
+++ b/third_party/external_files.bzl
@@ -550,6 +550,12 @@ def external_files():
         urls = ["https://storage.googleapis.com/mediapipe-assets/ssd_mobilenet_v1.tflite?generation=1661875947436302"],
     )
 
+    http_file(
+        name = "com_google_mediapipe_test_jpg",
+        sha256 = "798a12a466933842528d8438f553320eebe5137f02650f12dd68706a2f94fb4f",
+        urls = ["https://storage.googleapis.com/mediapipe-assets/test.jpg?generation=1664672140191116"],
+    )
+
     http_file(
         name = "com_google_mediapipe_test_model_add_op_tflite",
         sha256 = "298300ca8a9193b80ada1dca39d36f20bffeebde09e85385049b3bfe7be2272f",

From ab672938698ed328615b37c2453359b559be9b34 Mon Sep 17 00:00:00 2001
From: MediaPipe Team <mediapipe-team@google.com>
Date: Tue, 4 Oct 2022 15:38:58 -0700
Subject: [PATCH 5/5] Internal change

PiperOrigin-RevId: 478900121
---
 .../cc/vision/gesture_recognizer/proto/BUILD  | 10 +++++++
 .../gesture_embedder_graph_options.proto      | 30 +++++++++++++++++++
 2 files changed, 40 insertions(+)
 create mode 100644 mediapipe/tasks/cc/vision/gesture_recognizer/proto/gesture_embedder_graph_options.proto

diff --git a/mediapipe/tasks/cc/vision/gesture_recognizer/proto/BUILD b/mediapipe/tasks/cc/vision/gesture_recognizer/proto/BUILD
index cb6ec8289..7b5c65eab 100644
--- a/mediapipe/tasks/cc/vision/gesture_recognizer/proto/BUILD
+++ b/mediapipe/tasks/cc/vision/gesture_recognizer/proto/BUILD
@@ -20,6 +20,16 @@ package(default_visibility = [
 
 licenses(["notice"])
 
+mediapipe_proto_library(
+    name = "gesture_embedder_graph_options_proto",
+    srcs = ["gesture_embedder_graph_options.proto"],
+    deps = [
+        "//mediapipe/framework:calculator_options_proto",
+        "//mediapipe/framework:calculator_proto",
+        "//mediapipe/tasks/cc/core/proto:base_options_proto",
+    ],
+)
+
 mediapipe_proto_library(
     name = "hand_gesture_recognizer_graph_options_proto",
     srcs = ["hand_gesture_recognizer_graph_options.proto"],
diff --git a/mediapipe/tasks/cc/vision/gesture_recognizer/proto/gesture_embedder_graph_options.proto b/mediapipe/tasks/cc/vision/gesture_recognizer/proto/gesture_embedder_graph_options.proto
new file mode 100644
index 000000000..c12359eb3
--- /dev/null
+++ b/mediapipe/tasks/cc/vision/gesture_recognizer/proto/gesture_embedder_graph_options.proto
@@ -0,0 +1,30 @@
+/* Copyright 2022 The MediaPipe Authors. All Rights Reserved.
+
+Licensed under the Apache License, Version 2.0 (the "License");
+you may not use this file except in compliance with the License.
+You may obtain a copy of the License at
+
+    http://www.apache.org/licenses/LICENSE-2.0
+
+Unless required by applicable law or agreed to in writing, software
+distributed under the License is distributed on an "AS IS" BASIS,
+WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+See the License for the specific language governing permissions and
+limitations under the License.
+==============================================================================*/
+
+syntax = "proto2";
+
+package mediapipe.tasks.vision.gesture_recognizer.proto;
+
+import "mediapipe/framework/calculator.proto";
+import "mediapipe/tasks/cc/core/proto/base_options.proto";
+
+message GestureEmbedderGraphOptions {
+  extend mediapipe.CalculatorOptions {
+    optional GestureEmbedderGraphOptions ext = 478825422;
+  }
+  // Base options for configuring hand gesture recognition subgraph, such as
+  // specifying the TfLite model file with metadata, accelerator options, etc.
+  optional core.proto.BaseOptions base_options = 1;
+}