From: https://google.github.io/mediapipe/
python -m venv env
source env/bin/activate
pip install -r requirements.txt
For the iris example, put iris_landmark.tflite into models directory, by unpacking following zip file:
https://github.com/google/mediapipe/files/10012191/iris_landmark.zip
The facial expression example uses the trained weights from github.com/zengqunzhao/EfficientFace, but converted to tflite.
For the facial expression example download both models (fast and slow) into the models directory:
wget -P models https://rassibassi-mediapipedemos.s3.eu-central-1.amazonaws.com/efficient_face_model.tflite
wget -P models https://rassibassi-mediapipedemos.s3.eu-central-1.amazonaws.com/dlg_model.tflite
One of the following:
python facial_expression.py
python face_detection.py
python face_mesh.py
python hands.py
python head_posture.py
python holistic.py
python iris.py
python objectron.py
python pose.py
python selfie_segmentation.py
pose.py and iris.py include the possibility to process a video file instead of the webcam input stream. Run like this:
python iris.py -i /path/to/some/file/i-am-a-video-file.mp4
python pose.py -i /path/to/some/file/i-am-a-video-file.mp4
See example python pose.py for how to extract numpy array from the mediapipe landmark objects.