Invariant face recognition using a neural network based on the fringe-adjusted joint transform correlator

In this paper, we propose an optoelectronic two-layer neural network based on the fringe-adjusted joint transform correlator for invariant face recognition accommodating in-plane and out-of-plane 3-D distortions. The neural network is utilized in the training stage for a sequence of facial images and for supervised learning in order to create composite images that are invariant to 3-D distortions. The proposed technique is implemented by using the fringe-adjusted joint transform correlator. Simulation results are presented to verify the performance of the proposed technique. These results are then compared with those obtained using other techniques such as the synthetic discriminant function. (C) 2002 Society of Photo-Optical Instrumentation Engineers.