High-precision Immune Computation for Secure Face Recognition

The accuracy of face recognition is very important for its security in many applications, because wrong face recognition may cause such security problems as authorization. To increase the recognition rate in such face database as ORL, a face recognition algorithm should be good at minimizing the disturbances of facial pose, illumination and expression (PIE) to this recognition. In this paper, the improved clonal selection algorithm and diverse samples are designed. The improved clonal selection algorithm searches the most similar sample for unknown face image, according to the affinity between the unknown one and the most similar sample. The affinity is newly designed to improve the adaptive matching between the object and the samples. Compared with some state-of-the-art algorithms on the ORL face database, the proposed approach outperforms the other algorithms in the recognition rate, based on the experimental results.