Gabor Based Face Recognition with Dynamic Time Warping

Many approaches are proposed for face recognition from last three decades but they are challenging the problems like illumination, pose and scaling variations. In this paper a new methodology has been proposed for improving the efficiency of face recognition based on local feature extraction using Gabor Wavelets. The extracted features are classified using non linear matching algorithm like Dynamic Time Warping (DTW). DTW is a technique which is used to identify an optimal warp between two feature vectors. Based on the constraints of DTW it provides better accuracy than the existing methods with Euclidean distance. The proposed method has given 96.19% recognition rate on Grimace face database, 86.38% of recognition rate on ORL and 90.67% on Yale like standard bench mark face databases.