Illumination compensation method using edge-weakening guided image filter for face recognition

In face recognition, the self-quotient image (SQI) is a widely used preprocessing step to compensate for illumination. To improve SQI performance, a specially designed edge-weakening guided image filter (EWGIF) is proposed. An edge aware coefficient in the EWGIF makes it aware of edges and able to change the smoothing level automatically according to the image contents. Through the improved SQI, most of the face informatic areas are well preserved, while inconsequential parts are dropped. Experimental results on the extended Yale B and the Gray FERET databases show that face recognition accuracy is significantly improved with the EWGIF.