Repository contains implementation of Dlib face detector as well as Mediapipe Facemesh detector for head pose estimation.
Run pip install -r requirements.txt in the root directory of the project.
Code for the detector is located inside mediapipe_facemesh_head_pose_estimator.py
Runs real time head pose estimation on the webcam input and shows the output in a new window.
python mediapipe_facemesh_head_pose_estimator.py
Runs head pose estimation on the whole video and saves it to out_path
python mediapipe_facemesh_head_pose_estimator.py --video True --video_path "video_path" --out_path "out_path
Returns pitch, yaw roll for the provided image
python mediapipe_facemesh_head_pose_estimator.py --image True --image_path "image_path"
Code for the detector is located inside dlib_head_pose_estimator.py
Runs real time head pose estimation on the webcam input and shows the output in a new window.
Scale percent option (--scale_percent x) scales the image to x% of initial size.
python dlib_head_pose_estimator.py --scale_percent 30
Runs head pose estimation on the whole video and saves it to out_path.
Scale percent option (--scale_percent x) scales the image to x% of initial size.
python dlib_head_pose_estimator.py --video True --video_path "video_path" --out_path "out_path --scale_percent 30
Returns pitch, yaw roll for the provided image.
python dlib_head_pose_estimator.py --image True --image_path "image_path"