Mediapipe

* Added 3D Face Transform to [MediaPipe Face Mesh](https://solutions.mediapipe.dev/face_mesh)
* Enables [Face Effect Rendering](https://solutions.mediapipe.dev/face_mesh#effect-renderer)
* Demonstrated with an [Face Effect Example App](https://solutions.mediapipe.dev/face_mesh#face-effect-example)
* More info in [Google Developers Blog post](https://mediapipe.page.link/face-geometry-blog)
* Added [MediaPipe Models and Model Cards](https://solutions.mediapipe.dev/models) doc

* [MediaPipe Instant Motion Tracking](https://solutions.mediapipe.dev/instant_motion_tracking) that performs AR tracking without initialization or calibration.
* [Google Developers Blog post](https://mediapipe.page.link/instant-motion-tracking-blog)
* [MediaPipe BlazePose in Python](https://solutions.mediapipe.dev/pose#python) now uses `pip install mediapipe` instead of building from source. Also published a [Colab](https://mediapipe.page.link/mp-py-colab) example.
* [MediaPipe Iris](https://solutions.mediapipe.dev/iris) updated to output a set of 478 3D landmarks, including 468 face landmarks from [MediaPipe Face Mesh](https://solutions.mediapipe.dev/face_mesh), with those around the eyes further refined, and 10 additional iris landmarks appended at the end.

* [MediaPipe Pose](https://solutions.mediapipe.dev/pose) for upper-body pose tracking
* [Google AI blog post for full and upper body Blazepose](https://mediapipe.page.link/blazepose-blog)
* Preview of [Python support](https://google.github.io/mediapipe/getting_started/building_examples.html#python) that runs MediaPipe Pose in Python interpreter

* [MediaPipe Iris](https://solutions.mediapipe.dev/iris) for iris tracking and single-image depth-from-iris
* Fixed mirrored text rendering in iOS example apps
* [Google AI blog post on MediaPipe Iris](https://mediapipe.page.link/iris-blog)
* [Iris tracking web demo](https://viz.mediapipe.dev/demo/iris_tracking)
* [Depth estimation from Iris web demo](https://viz.mediapipe.dev/demo/iris_depth)

* Added support for automatic provisioning for [building iOS examples](https://google.github.io/mediapipe/getting_started/building_examples.html#ios)
* Fixed sample trace file (https://github.com/google/mediapipe/issues/849)

* [TfLiteTensorsToLandmarksCalculator](https://github.com/google/mediapipe/blob/master/mediapipe/calculators/tflite/tflite_tensors_to_landmarks_calculator.cc)
* It now always [normalizes Z coordinates by image width (as for X)](https://github.com/google/mediapipe/blob/e9fbe868e55fa23aaabc31f9f847c22287062850/mediapipe/calculators/tflite/tflite_tensors_to_landmarks_calculator.cc#L223) when producing NORM_LANDMARKS output, assuming a weak perspective projection camera model.
* The [normalize_z](https://github.com/google/mediapipe/blob/e9fbe868e55fa23aaabc31f9f847c22287062850/mediapipe/calculators/tflite/tflite_tensors_to_landmarks_calculator.proto#L53) option can be used to further normalize Z coordinates by an additional factor. For instance, [normalize_z: 0.4](https://github.com/google/mediapipe/blob/e9fbe868e55fa23aaabc31f9f847c22287062850/mediapipe/graphs/hand_tracking/subgraphs/hand_landmark_gpu.pbtxt#L149) is now used in hand tracking to better account for the Z coordinate distribution in the training data.
* Bug fixes
* Fixes issues on Ubuntu 20.04, resolving 820.
* Documentation update
* Added [instructions](https://google.github.io/mediapipe/getting_started/install.html#installing-on-debian-and-ubuntu) for building for Nvidia Jetson and Raspberry Pi devices with ARM Ubuntu.