My project requires face recognition based on the availability of limited features which mainly include eyes and the forehead region. But I am unable to find any good resources to perform the same. There are mostly resources dealing with face detection to face recognition using opencv but I have not been able to find any resource on how I can use eye detection and its feature extraction to perform face recognition with opencv.
Any guidance regarding this problem or suggestions of alternatives would be helpful. Thank you
My idea was to use virtual mask application on basic full face image and then to make standard face_recognition / dlib's face encodings with large 68-points model for further dlib face comparation. It works. Not 100%, especially on millions-of-images data set but better then without such approach. https://github.com/bohdansok/Face_Recognition