Face Manifold Analysis Based on LFA Features

Some research efforts have shown that face images possibly reside on a nonlinear sub-manifold. Based on the Laplacian Eigen map, Laplacianfaces method was proposed. Laplacianfaces explicitly considers the manifold structure of the face image. To avoid the singular problem, Laplacianfaces method first project the image vectors to PCA subspaces. PCA produces global non-topographic linear filters. In this paper, we propose a novel approach. LFA instead of PCA is applied contrast with Laplacianfaces. LFA can capture local characteristics with little lose of global information and present an effective low dimensional representation of images. By combining LFA and LPP, the new algorithm outperforms than Laplacianfacesand has explicit significance, which is shown by a series of experiments.