Illumination Compensation for Face Image Using Retinex Method Based on Bilateral Filtering

Illumination variation is one of the critical factors affecting face recognition rate. In the paper, we propose a novel Retinex method based on bilateral filtering to compensate the illumination for face image. The proposed method estimates the illumination using the bilateral filtering, and applies the Retinex theory to obtain the reflectance image which is illumination invariant. The experimental results validate the performance of the method.