Face Recognition under Complex Conditions

Human face recognition plays an important role in application such as human computer interface, video surveillance and face image database management. Automatic face recognition is a very challenging technique. Up to date, there are still substantial challenging problems which remain to be solved. This paper presents an automatic face recognition solution. At first, the AdaBoost method is used for face detection, then a feature extraction based on wavelet transform and KPCA is proposed. These features were fed up into support vector machine for recognition. Experimental results showed that the classifier which we trained is able to detect faces in the cases of multi-pose and multi-face under complex background. The method we proposed is superior to traditional PCA in the time of features extraction. Lots of tests showed that successful face recognition over a wide range of illuminstion, pose, and expression in images from database.