Abstract:

An identity authentication method based on adaptive LBP face recognition is proposed.In the identity feature recording phase,Harr face cascade classifier is first employed on face sample library to detect the face area,then PCA method is adopted to reduce dimensionality for face area,an adaptive  LBP face recognition algorithm is adopted to extract the feature values of face samples,and at last these feature values are trained to construct the face feature library.The face threshold feature library is constructed by multiscene face image library,threshold queue,and multithreshold matching of faces in the whole group. In the identity authentication phase,the LBP matching method is used between the face of the certificating user and face threshold feature in the face threshold feature library so as to determine the best LBP threshold.Using this threshold,the face of the certificating user is matched with the face feature in the face feature library in order to confirm the system permission for the certificating user.The experiments show that the identity authentication method based on adaptive LBP face recognition is of  higher recognition rate and robustness.   

Key words: adaptive LBP;face recognition;Harr face cascade classifier;PCA;authentication