Illumination compensation algorithm based on quadratic polynomial model

Face detection and recognition technologies have a wide range of applications, while varying illumination typically make these technologies perform poorly. The illumination variations heavily affect the effectiveness of face visualization and face recognition. The novelty in this paper is that quadratic polynomial model is applied to illumination compensation. Experimental results demonstrate that the proposed algorithm performs well not only in weakening illumination effect but also in improving the face recognition rates. From the experiments, we can find that face recognition rate can reach 99.17%, and the visualization effectiveness of images is greatly improved.