This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
I had limited success with OpticalFlow. With FaceDetect I plot maybe 13 points and use distance between eyes to define z. Very brittle and not commercial-grade, but may yield face rotation and scale. Brittle, due to lighting and obscured features.
Also OpenCV, but ~30MB and maybe Python: https://github.com/kurnianggoro/GSOC2017
... or FaceMesh (3MB) result read into OpenCV? https://blog.tensorflow.org/2020/03/face-and-hand-tracking-in-browser-with-mediapipe-and-tensorflowjs.html
