Currently, I'm working on CNNs for face point detection, and i'm thinking about using face point for face recognition. I know facial point detection is an important step in face recognition. But I don't know how to use it and google didn't return any expected result.
The only algorithm I know that using face keypoint is Elastic Bunch Graph Matching which using Gabor wavelet.
So. is there any other method? please list informations as many as you know.
Great thanks for your help :)
Face key-points are very useful for face recognition. although their relative distance change as soon as face pose changes. So the recognition algorithm which is based on face key-points distance will not perform well.
The alternative usage from face key-points are for face alignment and normalization, where all the faces will be aligned to the same position. there are numerous algorithm which use face key-points to pre-process the input faces into a normal position.
Another approach would be to use face key-points for local feature extraction. It means to extract some features from each face key-point. such methods are used in algorithms like EBGM and AAM.