Locality Preserving Discriminant Projection

Abstract

Face is the most powerful biometric as far as human recognition system is concerned which is not the case for machine vision. Face recognition by machine is yet incomplete due to adverse, unconstrained environment. Out of several attempts made in past few decades, subspace based methods appeared to be more accurate and robust. In the present proposal, a new subspace based method is developed. It preserves the local geometry of data points, here face images. In particular, it keeps the neighboring points which are from the same class close to each other and those from different classes far apart in the subspace. The first part can be seen as a variant of locality preserving projection (LPP) and the combination of both the parts is mentioned as locality preserving discriminant projection (LPDP). The performance of the proposed subspace based approach is compared with a few other contemporary approaches on some benchmark databases for face recognition. The current method seems to perform significantly better.