API for c++ ImageSegmenter to get labels

PiperOrigin-RevId: 516714139

mediapipe/tasks/cc/vision/image_segmenter/BUILD

@@ -25,6 +25,7 @@ cc_library(
     visibility = ["//visibility:public"],
     deps = [
         ":image_segmenter_graph",
+        "//mediapipe/framework:calculator_cc_proto",
         "//mediapipe/framework/api2:builder",
         "//mediapipe/framework/formats:image",
         "//mediapipe/framework/formats:rect_cc_proto",
@@ -34,10 +35,13 @@ cc_library(
         "//mediapipe/tasks/cc/vision/core:image_processing_options",
         "//mediapipe/tasks/cc/vision/core:running_mode",
         "//mediapipe/tasks/cc/vision/core:vision_task_api_factory",
+        "//mediapipe/tasks/cc/vision/image_segmenter/calculators:tensors_to_segmentation_calculator_cc_proto",
         "//mediapipe/tasks/cc/vision/image_segmenter/proto:image_segmenter_graph_options_cc_proto",
         "//mediapipe/tasks/cc/vision/image_segmenter/proto:segmenter_options_cc_proto",
+        "//mediapipe/util:label_map_cc_proto",
         "@com_google_absl//absl/memory",
         "@com_google_absl//absl/status:statusor",
+        "@com_google_absl//absl/strings:str_format",
         "@org_tensorflow//tensorflow/lite/kernels:builtin_ops",
     ],
 )

mediapipe/tasks/cc/vision/image_segmenter/image_segmenter.cc

@@ -15,15 +15,21 @@ limitations under the License.
 
 #include "mediapipe/tasks/cc/vision/image_segmenter/image_segmenter.h"
 
+#include <optional>
+
+#include "absl/strings/str_format.h"
 #include "mediapipe/framework/api2/builder.h"
+#include "mediapipe/framework/calculator.pb.h"
 #include "mediapipe/framework/formats/image.h"
 #include "mediapipe/framework/formats/rect.pb.h"
 #include "mediapipe/tasks/cc/core/utils.h"
 #include "mediapipe/tasks/cc/vision/core/image_processing_options.h"
 #include "mediapipe/tasks/cc/vision/core/running_mode.h"
 #include "mediapipe/tasks/cc/vision/core/vision_task_api_factory.h"
+#include "mediapipe/tasks/cc/vision/image_segmenter/calculators/tensors_to_segmentation_calculator.pb.h"
 #include "mediapipe/tasks/cc/vision/image_segmenter/proto/image_segmenter_graph_options.pb.h"
 #include "mediapipe/tasks/cc/vision/image_segmenter/proto/segmenter_options.pb.h"
+#include "mediapipe/util/label_map.pb.h"
 
 namespace mediapipe {
 namespace tasks {
@@ -112,6 +118,39 @@ ConvertImageSegmenterOptionsToProto(ImageSegmenterOptions* options) {
   return options_proto;
 }
 
+absl::StatusOr<std::vector<std::string>> GetLabelsFromGraphConfig(
+    const CalculatorGraphConfig& graph_config) {
+  bool found_tensor_to_segmentation_calculator = false;
+  std::vector<std::string> labels;
+  for (const auto& node : graph_config.node()) {
+    if (node.calculator() ==
+        "mediapipe.tasks.TensorsToSegmentationCalculator") {
+      if (!found_tensor_to_segmentation_calculator) {
+        found_tensor_to_segmentation_calculator = true;
+      } else {
+        return absl::Status(CreateStatusWithPayload(
+            absl::StatusCode::kFailedPrecondition,
+            "The graph has more than one "
+            "mediapipe.tasks.TensorsToSegmentationCalculator."));
+      }
+      TensorsToSegmentationCalculatorOptions options =
+          node.options().GetExtension(
+              TensorsToSegmentationCalculatorOptions::ext);
+      if (!options.label_items().empty()) {
+        for (int i = 0; i < options.label_items_size(); ++i) {
+          if (!options.label_items().contains(i)) {
+            return absl::Status(CreateStatusWithPayload(
+                absl::StatusCode::kFailedPrecondition,
+                absl::StrFormat("The lablemap have no expected key: %d.", i)));
+          }
+          labels.push_back(options.label_items().at(i).name());
+        }
+      }
+    }
+  }
+  return labels;
+}
+
 }  // namespace
 
 absl::StatusOr<std::unique_ptr<ImageSegmenter>> ImageSegmenter::Create(
@@ -140,13 +179,22 @@ absl::StatusOr<std::unique_ptr<ImageSegmenter>> ImageSegmenter::Create(
                               kMicroSecondsPerMilliSecond);
         };
   }
-  return core::VisionTaskApiFactory::Create<ImageSegmenter,
-                                            ImageSegmenterGraphOptionsProto>(
-      CreateGraphConfig(
-          std::move(options_proto),
-          options->running_mode == core::RunningMode::LIVE_STREAM),
-      std::move(options->base_options.op_resolver), options->running_mode,
-      std::move(packets_callback));
+
+  auto image_segmenter =
+      core::VisionTaskApiFactory::Create<ImageSegmenter,
+                                         ImageSegmenterGraphOptionsProto>(
+          CreateGraphConfig(
+              std::move(options_proto),
+              options->running_mode == core::RunningMode::LIVE_STREAM),
+          std::move(options->base_options.op_resolver), options->running_mode,
+          std::move(packets_callback));
+  if (!image_segmenter.ok()) {
+    return image_segmenter.status();
+  }
+  ASSIGN_OR_RETURN(
+      (*image_segmenter)->labels_,
+      GetLabelsFromGraphConfig((*image_segmenter)->runner_->GetGraphConfig()));
+  return image_segmenter;
 }
 
 absl::StatusOr<std::vector<Image>> ImageSegmenter::Segment(

mediapipe/tasks/cc/vision/image_segmenter/image_segmenter.h

@@ -189,6 +189,18 @@ class ImageSegmenter : tasks::vision::core::BaseVisionTaskApi {
 
   // Shuts down the ImageSegmenter when all works are done.
   absl::Status Close() { return runner_->Close(); }
+
+  // Get the category label list of the ImageSegmenter can recognize. For
+  // CATEGORY_MASK type, the index in the category mask corresponds to the
+  // category in the label list. For CONFIDENCE_MASK type, the output mask list
+  // at index corresponds to the category in the label list.
+  //
+  // If there is no labelmap provided in the model file, empty label list is
+  // returned.
+  std::vector<std::string> GetLabels() { return labels_; }
+
+ private:
+  std::vector<std::string> labels_;
 };
 
 }  // namespace image_segmenter

mediapipe/tasks/cc/vision/image_segmenter/image_segmenter_test.cc

@@ -15,6 +15,7 @@ limitations under the License.
 
 #include "mediapipe/tasks/cc/vision/image_segmenter/image_segmenter.h"
 
+#include <array>
 #include <cstdint>
 #include <memory>
 
@@ -71,6 +72,13 @@ constexpr float kGoldenMaskSimilarity = 0.98;
 // 20 means class index 2, etc.
 constexpr int kGoldenMaskMagnificationFactor = 10;
 
+constexpr std::array<absl::string_view, 21> kDeeplabLabelNames = {
+    "background", "aeroplane",    "bicycle", "bird",  "boat",
+    "bottle",     "bus",          "car",     "cat",   "chair",
+    "cow",        "dining table", "dog",     "horse", "motorbike",
+    "person",     "potted plant", "sheep",   "sofa",  "train",
+    "tv"};
+
 // Intentionally converting output into CV_8UC1 and then again into CV_32FC1
 // as expected outputs are stored in CV_8UC1, so this conversion allows to do
 // fair comparison.
@@ -244,6 +252,22 @@ TEST_F(CreateFromOptionsTest, FailsWithInputChannelOneModel) {
                         "channels = 3 or 4."));
 }
 
+TEST(GetLabelsTest, SucceedsWithLabelsInModel) {
+  auto options = std::make_unique<ImageSegmenterOptions>();
+  options->base_options.model_asset_path =
+      JoinPath("./", kTestDataDirectory, kDeeplabV3WithMetadata);
+  options->output_type = ImageSegmenterOptions::OutputType::CATEGORY_MASK;
+
+  MP_ASSERT_OK_AND_ASSIGN(std::unique_ptr<ImageSegmenter> segmenter,
+                          ImageSegmenter::Create(std::move(options)));
+  const auto& labels = segmenter->GetLabels();
+  ASSERT_FALSE(labels.empty());
+  ASSERT_EQ(labels.size(), kDeeplabLabelNames.size());
+  for (int i = 0; i < labels.size(); ++i) {
+    EXPECT_EQ(labels[i], kDeeplabLabelNames[i]);
+  }
+}
+
 class ImageModeTest : public tflite_shims::testing::Test {};
 
 TEST_F(ImageModeTest, SucceedsWithCategoryMask) {