Detection And Localization Of Face And Face Recognition Algorithm Based On Feature Fusion

Due to the development of technology and application requirements, as well as the diversification of recognition, people attach great attention to fingerprint identification and face recognition technology. Face recognition technology involves image processing, pattern recognition, artificial intelligence and so on, because of the support vector machine, face recognition has become one of the most popular researches in the field of computer vision. A face recognition system usually contains face detection, facial feature locatization, facial feature extraction and facial feature matching, however, the influence of face and external conditions become a barrier for fast and efficient face detection and feature extraction. The main research work of this paper is the face detection, facial feature location and face recognition based on support vector machine.Whether it can detect face rapidly and accurately, has great impact on the subsequent research of face recognition. In order to detect all the face in a image,firstly this paper thought the skin color detection to excluding large amounts of complex background of non-face, then define the face candidate regions. Besides, then through AdaBoost algorithm to detect the human face, to improve the accurate rate of face detection system and reduce the error rate. In addition, based on the AdaBoost algorithm of former research, this paper adds new Haar features and modify the weight of its update method, so that under the condition of the less weak classifier, the AdaBoost algorithm’s training speed more faster, and to prevent the excessive distribution phenomenon in the process of the training.For the detected images, this paper put forward the active shape model and active appearance model fusion algorithm for facial feature locatization. Using active shape model in positioning face peripheral contour, active shape model can get accurate localization result for face peripheral contour. Active active appearance model algorithm positions the internal organs feature points of face, because it considers the texture information. Combined with the two algorithms, makes the facial feature points positioning time shorter and precision much accurate.In order to identify different human faces, using the local texture feature of image and support vector machine algorithms, this paper proposes a new face recognition algorithm. Extracting the local binary pattern of image, as the input information of support vector machine to training features. Thus, it can solve the problems of imperfect extracting feature of local binary pattern features and ignoring the local structural features of support vector machine.