Facial feature extraction and recognition based on Curvelet transform and SVD

This paper proposed a new algorithm of facial feature extraction and recognition based on curvelet transform and SVD (singular value decomposition). Firstly, the method carries on the curvelet transform to the face images, extracts the curvelet energy features of low frequency and the high frequency, and then carries on the singular value compression and feature fusion to this component. Finally, by using the nearest neighbor classifier in the ORL person face database, the classified recognition was carried out so that to confirm the validity of the algorithm.