Research On Automatic Face Recognition

Speedy and efficient automatic authentications are more and more needed after the"9·11"event. As the inherent quality of human beings, the biological characteristics with self-stability and individual diversity so it becomes the most ideal basis of the authentication. Among the ways for authentication, recognition by human faces is the most natural and direct way. Compared to the other biological characteristics, it is more easily accepted by the users because it is direct, friendly and convenient. The technology for human face recognition became a challenging issue in the fields of pattern recognition and computer vision, and became a focus in recent years'studies.Up to now, there are many ways of human face recognition which have been raised, tested and even put into practical application. So we think that it is very necessary to summarize the research work in recent years. Based on the studies and researches of human face detection and location, and human face feature extraction and recognition, this article will summarize ways of human face detection and recognition, and point out the central issue of these two subjects. This article primarily studies new algorithms with better robustness, and makes performance evaluation and analysis of the referred algorithms from the perspectives of discerning rate, the complexity of calculation, resources taken up and practical application.