Using artificial immunity network for face verification

Biometrical systems are of the most interesting research subject matters in the last years. Face biometrics is noteworthy one because of its simple accessibility, easy usage and the ability of better acceptance by persons. The process of facial recognition includes these phases: Pre-processing of images, extracting important properties of the face, and finally, the classification of these properties. There are many researches carried out in this area, each of which employed different methods for mentioned phases. According to the previous applications of the methods, which have been done by artificial immune network, and to its relatively good results in optimization problems, machine learning, pattern recognition, data search, data clustering and so on, in this research facial verification through classification by artificial immune Network (aiNET) has been surveyed. In this article, databank Yale has been used and the statistical properties such as maximum, minimum, variance and energy of wavelet coefficients in different compositions have been examined. In order to validation, we have used the Cross Validation method that its best results in the case of using the ten-fold or leave one out method, were FAR=2.1 %, FRR=0.9%, and EER=1.8%.