Face recognition based on feature fusion

A face recognition method based on the feature fusion of the local and global features is proposed. Each face image is divided in two vertical, and upper sub-images from the same position consruct a new training sub-set, a set of feature spaces can be obtained by training the face image set and sub-image set, based on Principal Component Analysis. Last, k-nearest neighbor classifier is used to recognize different faces from the ORL face database. Experimental results show that the feature fusion method improved the recognition rate effectively in comparison with the traditional PCA method, The best accuracy rate can reach 90%.