Invariant face recognition by Gabor wavelets and neural network matching

This paper presents a model-based face recognition approach that uses a hierarchical Gabor wavelet representation and neural network matching. Local features of grey level images are extracted by multiresolutions of Gabor wavelets, which are scaled and rotated versions of each other. The Gabor wavelet representation is use in a innovative neural network matching approach that can provide robust recognition. Neural network matching between a model and a input image is to find out the exact correspondence of local features and to map the model to the input image based on local similarity and neighborhood grouping of local features. The results on face recognition are presented, where the objects undergo rotation, translation, local distortions, and deformation caused by facial expression.