Face illumination invariant feature extraction based on edge detection operator
Wenyi Xiong, Xiangfei Nie, Xing Zou, Zhijun Yang, Xue He · 2017
A new method of face illumination invariant feature extraction based on edge detection operator is proposed in this paper. The input face image is calculated by edge enhancement with edge operator. And then, original face image is divided by the sharpened image to eliminate the illumination component to get the quotient image which is the illumination invariant feature of face image. Finally, the arctangent transform is employed to do nonlinear compression. The experiment results of face recognition on Extended Yale B and CMU PIE face database show that the method is robust to face non-uniform illumination.