Distortion-invariant face recognition using multiple phase-shifted reference-based joint transform correlation technique

Abstract

We have developed a novel face recognition technique utilizing optical joint transform correlation (JTC) technique which provides with a number of salient features as compared to similar other digital techniques, including fast operation, simple architecture and capability of updating the reference image in real time. The proposed technique incorporates a synthetic discriminant function (SDF) of the target face estimated from a set of different training faces to make the face recognition performance invariant to noise and distortion. The technique then involves four different phase-shifted versions of the same SDF reference face, which are individually joint transform correlated with the given input scene with unknown faces and other objects. Appropriate combination of correlation signals yields a single cross-correlation peak corresponding to each potential face image. The technique also involves a fringe-adjusted filter to generate a delta-like correlation peak with high discrimination between the target face and the non-target face and background objects. Performance of the proposed face recognition technique is investigated through computer simulation where it is observed to be efficient and successful in different complex environments.