extract_face_bounds.txt

usage: extract_face_bounds.py [-h] --inframes INFRAMES --outbounds OUTBOUNDS
                              [--method METHOD] [--head-focus]

Get the bounding box of the face throughout a video

optional arguments:
  -h, --help            show this help message and exit
  --inframes INFRAMES, --invideo INFRAMES
                        Path to a video or directory showing a sign language
                        interpreter. Hence, we assume that there is a face
                        always present and visible.
  --outbounds OUTBOUNDS
                        Path for a JSON output file: a JSON structure
                        containing the pixel-coordinates of the smallest
                        rectangle containing the face of the person throughout
                        the whole video. The rectangle must hold the same
                        proportions of the original video (e.g.: 4:3, 16:9).
                        Output has the format: { "x": int, "y": int, "width":
                        int, "height": int}.
  --method METHOD       Select the framework used to extract the face
                        boundaries. Possible values: "mediapipe", "mtcnn"
  --head-focus          Used only with the "mediapipe" method. Before trying
                        to recognize the face, try to recognize the head zone
                        of a full body. Useful when the face is too small but
                        the body is visible. However, body recognition is much
                        slower.