cassdlcm/mediapipe
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
1,571 Commits 10 Branches 56 Tags 829 MiB
367ccbfdf3
Commit Graph

1571 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
MediaPipe Team
ffea85e470 Internal change 
PiperOrigin-RevId: 516179167
 2023-03-13 05:13:01 -07:00
kinaryml
89be4c7b64 Added some files for the face landmarker implementation 2023-03-12 16:09:04 -07:00
MediaPipe Team
131be2169a Add FaceDetector Java API 
PiperOrigin-RevId: 515913662
 2023-03-11 13:14:47 -08:00
MediaPipe Team
296ee33be5 Add FaceLandmarker C++ API 
PiperOrigin-RevId: 515912777
 2023-03-11 13:05:21 -08:00
MediaPipe Team
c94de4032d Fix preprocess Callable typing 
PiperOrigin-RevId: 515818356
 2023-03-10 21:50:46 -08:00
Jiuqiang Tang
3e8fd58400 Make createAudioRecord a class method not a static method. 
PiperOrigin-RevId: 515740313
 2023-03-10 14:34:32 -08:00
Sebastian Schmidt
db779ba78f Use drishti_proto_library for libraries in mediapipe/gpu 
PiperOrigin-RevId: 515708036
 2023-03-10 12:27:06 -08:00
MediaPipe Team
c9bd4f5957 Internal change 
PiperOrigin-RevId: 515706419
 2023-03-10 12:22:27 -08:00
MediaPipe Team
c3a32d76be Update face geometry proto java package name. 
PiperOrigin-RevId: 515696170
 2023-03-10 11:45:35 -08:00
MediaPipe Team
296c343332 Revise the Halide Bazel build rules to be cleaner and more correct. The initial version we landed happened to work (at least in most cases) but had known workarounds in it. AFAICT, the issue seemed to be that the MediaPipe config_setting() values we were using didn't always resolve to the right thing when detecting the host cpu (vs the compile-target cpu), which is critical for Halide. I fixed this by avoiding the use of the MediaPipe config_settings entirely, using instead the public Bazel settings hosted at https://github.com/bazelbuild/platforms. 
This revision builds and runs //mediapipe/util/frame_buffer:all correctly on my mac x86-64 laptop; I haven't yet attempted to build or test on any other platform, so there may well still be glitches, but I think this is more fundamentally sound than what we had before.

PiperOrigin-RevId: 515682507
 2023-03-10 10:59:17 -08:00
MediaPipe Team
05b505c8e2 Introduce api to disable service default initialization. 
PiperOrigin-RevId: 515501608
 2023-03-09 18:57:39 -08:00
Sebastian Schmidt
ef4a8cde42 Solve iOS build error for gpu_buffer.cc 
PiperOrigin-RevId: 515473643
 2023-03-09 16:28:58 -08:00
MediaPipe Team
2d8f937913 Improve docstring of image classifier model spec. 
PiperOrigin-RevId: 515466722
 2023-03-09 16:06:40 -08:00
Sebastian Schmidt
517e997179 Solve Linking error for Hello World iOS example 
PiperOrigin-RevId: 515466634
 2023-03-09 16:02:07 -08:00
MediaPipe Team
c2a69ab476 Update ImageFrameToGpuBufferCalculator to use api2 and GpuBuffer conversions 
PiperOrigin-RevId: 515407159
 2023-03-09 12:21:00 -08:00
MediaPipe Team
5daf58009d internal 
PiperOrigin-RevId: 515392932
 2023-03-09 11:27:11 -08:00
Jiuqiang Tang
5bd6a7082a Add requiredInputBufferSize as an input argument of createAudioRecord. 
PiperOrigin-RevId: 515374407
 2023-03-09 10:30:49 -08:00
MediaPipe Team
563b193bca Bump Halide version from 14.0.0 to 15.0.0 and add MacOS Halide dependency 
PiperOrigin-RevId: 515362310
 2023-03-09 09:54:19 -08:00
MediaPipe Team
2c64421471 Fix minor typos in MediaPipe synchronization description. 
PiperOrigin-RevId: 515362264
 2023-03-09 09:49:23 -08:00
MediaPipe Team
39e2c8351f Add build system for Halide and expose FrameBufferUtils. 
PiperOrigin-RevId: 515304264
 2023-03-09 05:06:18 -08:00
kinaryml
f48909cab6 Fixed score's data type 2023-03-09 02:13:34 -08:00
kinaryml
24114ec2fe Updated comment in test 2023-03-09 01:41:42 -08:00
kinaryml
022838a7f3 Added Face Detector implementation and tests 2023-03-09 01:36:39 -08:00
Sebastian Schmidt
5398b8881d Internal change 
PiperOrigin-RevId: 515225596
 2023-03-08 21:04:55 -08:00
MediaPipe Team
c12eae229f Internal change 
PiperOrigin-RevId: 515187906
 2023-03-08 17:36:04 -08:00
MediaPipe Team
8d9f627fd9 Internal change 
PiperOrigin-RevId: 515187139
 2023-03-08 17:32:21 -08:00
MediaPipe Team
9f1f4273d0 Support hair segmentation model in image segmenter 
PiperOrigin-RevId: 515151150
 2023-03-08 15:01:14 -08:00
MediaPipe Team
2fb62e4c29 Add volume_gain_db option to TensorsToAudioCalculator. 
PiperOrigin-RevId: 515150234
 2023-03-08 14:57:29 -08:00
Jiuqiang Tang
253a5b477e Implement face stylizer graph and its C++ API. 
PiperOrigin-RevId: 515139282
 2023-03-08 14:18:08 -08:00
Kris Tonthat
0a60c67667 Redirect "Getting started pages" from MediaPipe.dev to developers.google.com/mediapipe 
PiperOrigin-RevId: 515137109
 2023-03-08 14:10:45 -08:00
Yuqi Li
6f8c30abdd nit: update the description format in object_detector.MetadataWriter. 
PiperOrigin-RevId: 515113799
 2023-03-08 12:46:08 -08:00
MediaPipe Team
b8917ad31f Add FaceDetectorResult 
PiperOrigin-RevId: 515104977
 2023-03-08 12:10:13 -08:00
Sebastian Schmidt
09f63cbbe0 Internal change 
PiperOrigin-RevId: 515094293
 2023-03-08 11:33:38 -08:00
Jiuqiang Tang
ddc535e705 Add DftTensorFormat To TensorsToAudioCalculatorOptions. 
PiperOrigin-RevId: 515077766
 2023-03-08 10:38:40 -08:00
Hadon Nash
16c2e32a0d Internal change 
PiperOrigin-RevId: 515044500
 2023-03-08 08:29:53 -08:00
MediaPipe Team
13d7e1a7ba Changes the visibility of ImageTensorSpecs to public to allow clients use GetInputImageTensorSpecs(). 
PiperOrigin-RevId: 514853148
 2023-03-07 15:16:40 -08:00
MediaPipe Team
225dd36a63 Internal change 
PiperOrigin-RevId: 514835354
 2023-03-07 14:11:47 -08:00
Yuqi Li
46b5c4012d add custom metadata in metadata_schema. 
PiperOrigin-RevId: 514768756
 2023-03-07 10:15:09 -08:00
MediaPipe Team
414031867e Internal change 
PiperOrigin-RevId: 514763674
 2023-03-07 10:00:10 -08:00
Jiuqiang Tang
dbd6d72696 Add GetInputImageTensorSpecs into BaseVisionTaskApi for tasks api users to get input image tensor specifications. 
PiperOrigin-RevId: 514650593
 2023-03-07 00:44:07 -08:00
MediaPipe Team
2f2a74da6a Internal MediaPipe Tasks change. 
PiperOrigin-RevId: 514637484
 2023-03-06 23:17:35 -08:00
MediaPipe Team
c1b460920c normalized keypoint container. Detection adds optional keypoint list. 
PiperOrigin-RevId: 514624925
 2023-03-06 22:02:59 -08:00
MediaPipe Team
e4ac1a10c2 TensorsToSegmentation web postprocessing on GPU: softmax activation 
PiperOrigin-RevId: 514589911
 2023-03-06 19:08:53 -08:00
MediaPipe Team
bd9a2ee1fc Internal change 
PiperOrigin-RevId: 514580892
 2023-03-06 18:13:50 -08:00
MediaPipe Team
0337c7f52f Fix typo in ImageSegmenter.java 
PiperOrigin-RevId: 514534013
 2023-03-06 14:51:49 -08:00
MediaPipe Team
f15244997f Internal change 
PiperOrigin-RevId: 514498364
 2023-03-06 12:58:01 -08:00
Jiuqiang Tang
77fcaa9597 Allow TensorsToAudioCalculator to add and overlay adjacent windows. 
PiperOrigin-RevId: 514483756
 2023-03-06 12:08:57 -08:00
MediaPipe Team
c64f83fff4 Internal change 
PiperOrigin-RevId: 514452059
 2023-03-06 10:29:29 -08:00
MediaPipe Team
6c68d4c358 Internal change 
PiperOrigin-RevId: 514421618
 2023-03-06 08:42:19 -08:00
MediaPipe Team
96b2958f18 Add ViewProvider for ImageFrame in GpuBufferStorageYuvImage. 
PiperOrigin-RevId: 514344066
 2023-03-06 07:53:02 -08:00