Research On Face Detection And Recognition Algorithms

Face detection and recognition technology has been used in safety accessing, visual detection, content search and new man-machine interface areas. And now the research on face detection and recognition has become one of the most active research in Image Processing, Pattern Recognition and Artificial Intelligence area and so on.At first, we research recent face detection algorithms and use the machine learning algorithm—AdaBoost to detect face. We describe the face feature and extend the feature form. After deeply researching the weak classifier, single stage classifier and cascade stage classifier are carried out. The cascade stage classifier improves, whose aim is to make training time reduced. The algorithm gets good results in present condition. And the false alarms can be discarded.Then we research the method of feature extraction including SVD extraction and Gabor feature. Since Gabor feature is robust to illumination change, we use it for feature extraction. But all of the features which put into recognition can bring"dimension disaster"and the method which distributes each feature the same weight is not benefit for face recognition. We get the distinctive features through AdaBoost algorithm.Finally, features are put into SVM classifier for face recognition as SVM has strong ability of data division. The proposed algorithm can achieve well results and demonstrate it is robust to pose change and illumination change in some degree.