Ask Your Question

Revision history [back]

Actually its better to use model that is closest to our dataset to detect eyes and faces and so on, these xml files are suitable for datasets that are similar to their train dataset, as berak mention its good to use Other libraries or impelemntation . FaceX ( is one of the implementations that I suggested its good enough for detecting perspective face 's landmarks how ever it can't detect side faces. Dlib( seems good I tried hard to use it in visual studio , but i couldn't use it.