The Research And Application Of Face Recognition Technology

As one of the most successful applications of image analysis and understanding, face recognition has recently received significant attention, especially during the few years. However, the way of thoroughly solving the problem in this field is still far away, it also remains an active research field nowadays.This paper researches the new method of the technology of face detection which is based on the complexion model. By the study, the paper puts forward a face detection method based on the skin color and the position of eyes. First use the difference image to get an image that exist a man. Then get a face based on the skin color and the shape feature. Next the location of eyes and based on the location, find the top, bottom, left and right boundaries of face accurately. By this way, a face is detected.This paper also researches the new method of the technology of face recognition which is based on the integration of Eigenface, singular value decomposition, wavelet theory, BP Neural Network, and Hidden Marko Models. By the study, the paper puts forward an improved face recognition arithmetic based on SVM. Use a sampling window to sample from top to bottom. It can overlap. Count the strange values for all the matrices, compress the dimension, and get "k" maximal strange values of every matrix. These values are acted as the features of the image to recognize. Proved by experiment, the recognition rate is higher.Last, the paper implements a check on work attendance system prototype adopting face recognition technology. The algorithm combines the improved strange value decomposition (SVD) and classifying method based on minimal distance to recognize face. First based on the difference image and the skin color the face is detected and localized, then features are extracted by using improved SVD, last the features is classified by minimal distance. In order to enhance the robustness of the system, we use light compensation in the face detection and by fixing the position of eyes, we fix the face, it improves the veracity according to the face module. We do geometric standardization and gray-level standardization in face recognition, too. The system runs well and the effect is satisfying.The paper is composed of seven sections. The first section briefly recommends the relative research achievements and theories in face detection and recognition. The section two describes the relative theories in the fields of technology of face detection and the adopted arithmetic in this paper. The section three, section four and section five respectively describe the relative theories of principal component analysis, BP Neural network, Hidden Marko Models and the adopted arithmetic in this paper. The section three describes an improved face recognition arithmetic based on SVM. The last section designs a check on work attendance system adopting face recognition technology.