Abstract

In this paper, an efficient face recognition method based on fast discrete Curvelet\ntransform via wrapping is proposed. Feature vector is generated using the curvelet\ntransform coefficients of low frequency and high frequency with different scale and\norientation. Further dimensionality reduction of feature vector, principal component\nanalysis are used on feature space. Distance classifier is used to match the feature vector\ncoefficients of training dataset and testing image.The experiments are carried out for face\nrecognition using Face94 and IIT_Kanpur database.