Feature extraction based on fuzzy bidirectional maximum margin criterion for face recognition

This paper proposes a new method for feature extraction and recognition, namely, the fuzzy bidirectional maximum margin criterion (FBMMC) based on the maximum margin criterion and fuzzy set theory. In FBMMC, a membership grade matrix is calculated by the FKNN algorithm, and then the membership grades are incorporated into the formulation of the scatter matrices to formulate the row and column directional fuzzy scatter matrices. Using the fuzzy scatter matrix, FBMMC performs fuzzy image MMC in the row and column directions simultaneously. Experiments on the ORL and Yale face databases show the effectiveness of the proposed FBMMC based feature extraction method.