A novel maximum margin neighborhood preserving embedding for face recognition

This paper presents a novel supervised linear dimensionality reduction approach called maximum margin neighborhood preserving embedding (MMNPE). The central idea is to modify the neighborhood preserving embedding by maximizing the maximum margin distance while preserving the geometric structure of the manifold. Experimental results conducted on the ORL database, the Yale database and the VALID face database indicate the effectiveness of the proposed MMNPE.

Research highlights
► The proposed algorithm encodes discriminant information into LLE.
► Orthogonalization is used to obtain orthogonal basis vectors.
► In order to show its strength, face recognition is chosen as the primary application.