Face Recognition Using Gabor Volume Based Local Binary Pattern

The image space, scale and orientation domains can give valuable clues not seen in either individual of the domains. First we decomposed the face image into different orientation and scale by Gabor filter. Second, we combine local binary pattern analysis with Gabor. It gives a good face representation for recognition. Then we classify in discriminant based up on median histogram distance. The face images are usually affected by different expression, poses, illuminations and occlusions. Although the existing methods perform well under certain conditions, occlusions are still remaining as challenging problem. The proposed method median histogram distance deals this problem. This way, informations from different domains are explored to give a good face representation for recognition. Extensive experimental results on Yale database shows the significant advantages of the method. Index Terms Effective Gabor Volume Based Local Binary Pattern (E-GV-LBP), Face recognition, Gabor, Local Binary Pattern (LBP).