Constraint Shape Model Using Edge Constraint and Gabor Wavelet Based Search

Constraint Shape Model is proposed to extract facial feature using two different search methods for contour points and control points individually. In the proposed algorithm, salient facial features, such as the eyes and the mouth, are first localized and utilized to initialize the shape model and provide region constraints on iterative shape searching. For the landmarks on the face contour, the edge intensity is exploited to construct better local texture matching models. Moreover, for control points, the proposed Gabor wavelet based method is used to search it by multi-frequency strategy. To test the proposed approaches, on a database containing 500 labeled face images, experiments are conducted, which shows that the proposed method performs significantly better in terms of a deliberate performance evaluation method. The proposed method can be easily used to other texture objects, which is robust to variations in illumination and facial expression.