Author : Soju Ravi K

The face images are usually affected by different expressions, positions, occlusions and illuminations. The difference of face image from the same person could be different at positions. The major problem is to extract fair and discriminate features which make the intra person face images compact and provide large margin among the images of different persons. Since individual domains such as space, scale and orientation cannot provide sufficient information to discriminate the face images, a combined information extracted from these domains can provide rich information for representation of face images. Using Gabor filters, images are decomposed into different scale and orientation response. Local Binary Pattern analysis is performed to describe the neighboring relationship of the responses in these domains. Thus the information explored from the domains can give immense clue for face representation

Keywords: Face recognition, Gabor volume based local binary pattern (GV-LBP) ,Gabor volume representation, local binary pattern

Article published in International Journal of Current  Engineering  and Technology, Vol.3,No.5 (Dec- 2013)