A face recognition system on distributed evolutionary computing using on-line GA

Abstract

Although there is much research on face recognition, there now exists some limitations especially in illumination and pose. This paper addresses a novel framework to prevail over the illumination barrier and a robust vision system. The key ideas of this paper are distributed evolutionary computing and on-line GA that is the combining concept of context-awareness and genetic algorithm. This research implements Fuzzy ART that carries out the context-awareness, modeling, and identification for the context environment and the system can also distinguish changing environments. On-line GA stores the experiences to make context knowledge that is used for on-line adaptation. Finally, supervised learning is applied to carry on recognition experiments. Experimental results on FERET data set show that On-line GA based face recognition performance is significantly benefited over the application of existing GA classification