Face recognition using curvelet and selective PCA
Hanjun Huo, Song E-nuo · 2010
A second generation curvelet transform and selective PCA based face recognition method is proposed in this paper. Curvelet transform has better directional and edge representation abilities than widely used wavelet transform. It can provide better facial features for face recognition. The image is decomposed into its curvelet subbands at first, and distinctive curvelet detailed subbands are selected and cascaded to form a Cascaded-curveface, then PCA(principal component analysis) is applied on the Cascaded-curveface in order to create a representative feature. The experiments on ORL face database indicate that the propose method is of higher recognition rate.