Face Detection Of Eyes Grads Feature And Application Of Fisher Face Recognition

Biometrics is a kind of science of using individual personal characteristics to verify identity. It represents a new personal identification method based on a unique, highly reliable and steady physical characteristic. It is the safest manner of personal identification. Human face is the universal pattern of human being's vision. The visual information that face represents has important effect and meaning in society intercourse. Although a lot of arithmetic of face detection and recognition has come forth, the recognition rate, the recognition velocity and the robustness of arithmetic etc are not ideal enough yet. So, this thesis researches face detection and recognition deeply. The main works are as following. 1. Describes physical characteristic and its advantages, characteristics, the direction of development. Mainly introduces research content, method, application scope of face detection and recognition. 2. Aims at the problem of face detection in image with complex background, this thesis represents a method of face detection based on feature of eyes. The experimental results prove that this method can detect face with any position and various sizes in complex background rapidly and accurately. Furthermore, it can perform well for multiple faces image. 3. Introduces K-L transform and Fisher linear discriminating method in detail. Since illumination condition effects recognition ratio for human face greatly, while average gray value of every face image is different, we must standardize gray value of face images. This paper improves the PCA so that it can standardize gray value of face images; then combines the improved PCA and Fisher linear discrimination to recognize human face. We experimented on ORL face database, the experimental result proves that recognition ratio is higher evidently. Through studying the face detection and recognition technique, this thesis presents a method of face detection based on eyes feature and improves the PCA to recognize human face. The experimental result proves that the two methods are good.