Access Control System Face Recognition Technology. Heilongjiang Provincial Hospital

Face recognition technology is always being the focus of the research of Pattern Recognition and Artificial Intelligence. And it is standout among all the kinds of biometric-based recognition technology. It has been widely used in access control, e-commerce and video surveillance and other fields. The core problem of face recognition is how to extract effective facial features, which will affect the classification results, speed and recognition performance directly.In this paper, we design the access control system based on face recognition in hospital, including its hardware and software system. Our system is based on Linux embedded system, in which set ARM as the front video capture and simple face detection function, then it will send the pictures to the PC in background to complete the process of face recognition. For the limit of storage space and computing speed, front embedded microprocessor platform can not complete the process of training and classifying. So we should complete it on PC in background.We also discuss the mainstream of face recognition algorithms: algorithm based on Gabor wavelet transform. It can extract the features of face images more efficient than Fourier transform via a series of experiments. PCA is a traditional algorithm of face recognition and efficient and"Eigenface"is its practical application.On this basis, we proposed a fusion method based on Gabor and PCA method for face recognition. At first, we can extract more features via Gabor wavelet transform and PCA method will contribute to reduce the dimension of Gabor feature vectors. And it has been proved to be more efficient in the rate and speed of recognition.