Face Recognition by Integrating Chin Outline

Face recognition, one of the biometric methods identifying individuals by the features of face, has attracted increasing interest and research activities over the last decade. In this paper, we proposed a two-stage face recognition scheme inspired from multiresolution perspective. The lower resolution information is derived from the chin shape, whereas the higher resolution characteristics are extracted from the facial features. The chin shape information is characterized by the product of curvature and face length and the ratio of face width to face length, which are scale-independent. The facial features are extracted from Gabor faces followed by 2D-PCA for condensing features. A serial experiments were conducted to assess the system performance. The experimental results reveal that our system can effectively recognize faces and the chin shape information indeed can enhance the face recognition performance.