Releases: google-ai-edge/mediapipe
Releases · google-ai-edge/mediapipe
MediaPipe Android Solutions and Hand World Landmarks
- MediaPipe Hands, Face Detection, and Face Mesh Android Solutions are now available in Google's Maven Repository.
- The Android Studio example project is available in mediapipe/examples/android/solutions.
- MediaPipe Hands models are updated.
- MediaPipe Hands now supports outputting world landmarks in world coordinates.
MediaPipe Dependencies
- MediaPipe Python wheels are now supporting Python 3.10.
- The MediaPipe dependency library protobuf, tensorflow, cere solver, pybind, and apple support are updated.
- The recommended Bazel version is updated to 4.2.1.
- The recommended Android SDK and NDK versions are updated.
Additional options in MediaPipe Face Mesh and Holistic
- MediaPipe Face Mesh
- Added an
refine_landmarks
option in the Solution APIs to further improve landmarks around eyes and lips, and output additional landmarks around the irises. - Released the Attention Mesh ML model that enables
refine_landmarks
.
- Added an
- MediaPipe Holistic
- Added the
enable_segmentation
andsmooth_segmentation
option in the Solution APIs, previously only available in MediaPipe Pose.
- Added the
Merge PRs. Thank you to all contributors!
Merged #1969, #2039, #2101, #2136, #2154, #2201, #2291, #2292, #2347, #2350, #2450, #2462, and #2464.
A huge thank you to all contributors:
@Abduttayyeb
@brettkoonce
@chr0nikler
@daniel13520cs
@gabrielsanchez
@GantMan
@GzuPark
@homuler
@magamig
@PeterPocsi
@tomhsiao1260
@yuripourre
MediaPipe Pose Segmentation and Android Solution APIs
- MediaPipe Pose now outputs "segmentation_mask" when
enable_segmentation
is set. - MediaPipe Python FaceMesh, Hands, Pose, and Holistic have new drawing styles. Please refer to the documentation and the Colab examples for more details.
- MediaPipe now offers Android FaceMesh and Hands solution APIs (currently in Alpha). The maven artifacts are available in Google's Maven Repository. To try the example apps, please import the solution example Android Studio project on either Linux, macOS, and Windows.
3D world landmarks in MediaPipe Pose and full-range model in MediaPipe Face Detection
- MediaPipe Pose (and Holistic) now also outputs estimated real-world 3D coordinates of pose landmarks (in meters with the origin at center of hips)
- MediaPipe Face Detection now supports a "model_selection" option to switch between short-range and full-range models
MediaPipe Selfie Segmentation
- Released MediaPipe Selfie Segmentation
- JavaScript Solution API w/ a CodePen example
- Python Solution API w/ a Colab example
- Android, iOS and desktop example apps
Model update in MediaPipe Pose and MediaPipe Holistic
- Better pose landmark accuracy with the updated pose detection and pose landmark model
- Added a
lite
and aheavy
version of the pose landmark model, in addition to the existingfull
version (also updated and improved) - Added a
model_complexity
config option in MediaPipe Pose and MediaPipe Holistic Solution APIs to select across the 3 model versions - Removed the
upper_body_only
option in MediaPipe Pose and MediaPipe Holistic, as the standard model now already handles upper-body-only use cases well
Move minimum bazel version to 3.7.2
- Move up minimum bazel version to 3.7.2, and encourage the use of Bazelisk in Installation.
- Move up TensorFlow dependency.
- MediaPipe Objectron desktop/C++ example.
MediaPipe Pose with 3D landmarks, and more Python & JavaScript Solution APIs
- MediaPipe Pose
- New pose landmark model with Z coordinates (BlazePose GHUM 3D), generated from synthetic data
- Added tutorials and Colabs for Pose Classification
- MediaPipe Objectron
- Added Python Solution API
- MediaPipe Face Detection
- Added Python Solution API and JavaScript Solution API
- MediaPipe Face Mesh
- Face Geometry Module (code) now supports face-detection input (to generate a face geometry from a detection), in addition to the existing support for face-landmark input
- Image: a new multi-backend image data container:
- Covering both CPU and GPU storage
- Currently used in ImageToTensorCalculator and Face Detection Module
- Companion utility ToImageCalculator and FromImageCalculator
Experimental new framework APIs
- Released experimental new framework APIs
- Updated MediaPipe Objectron to support simultaneous detection of multiple objects
- Bug fixes
- Fixed a crash in the MediaPipe Iris iOS example app
- Fixed memory leak in Android example apps (in ExternalTextureConverter)
- Fixed a file read mode issue in pose_landmark_model_loader.pbtxt.