Robust face recognition under the polar coordinate system

Abstract

In this paper, we propose a novel method for face recognition which uses the polar coordinate system instead of the conventional cartesian coordinate system. Among the central area of a face, we select a point as a pole and make a polar image of a face by evenly sampling pixels in each direction of 360 degrees around the pole. The polar coordinate system delineates near-pole area more vividly than the area far from the pole. Therefore, a polar face image can achieve more vivid representation of the important central facial regions compared to the conventional cartesian face image. To cope with a small amount of rotation and illumination change in a frontal face image, we suggest a method based on vertical symmetry of a face. In addition, we also deal with a scaling problem by using a correlation-based method. Experimental results show that the proposed methods enhance classification performance in the frontal face recognition problem.