Training support vector machines for video based face recognition

In this paper the problem of training Support Vector Machines (SVMs) for video based face recognition is presented Faces as training samples are automatically extracted from input video sequence by multiple related template matching and normalized both in geometry via a)fine transformation based on corresponding facial feature points detected in the Sobel convolved face regions and in gray level distribution via linear transformation to the same average and squared difference. Two different strategies for q-class face recognition problems with SVMs are discussed both for ensemble face feature set and for PCA compressed feature set. The permance of a prototype system based on this technology oiler 100 clients is reported to demonstrate its great potentials in future.