A Passion Back Propagation Neural Network Approach for Face Recognition

Face recognition is a task of automatically identifying or verifying a person from a stored image. Even though a number of techniques have been developed for recognition human faces, none of them considered the way how human classify faces. This paper presents a novel approach for face recognition, which classifies human faces by simulating the behavior of human’s “glance” and face “affinity”. A new Passion back propagation (PBP) learning algorithm is developed based on two essential emotions (anxiety, certainty). Pattern Mean method is employed for extracting features which mimics the human Passion judgments based on general impressions rather than precise details of the objects. The performance of the proposed algorithm is tested using 400 facial images from the ORL face database. Simulation results show that the proposed emotion based BP algorithm (PBP) yields faster learning rate and takes lesser recognition time when compared with conventional BP algorithm and other approaches reported in the literature.