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.
face recognition LTV model quadratic polynomial model illumination compensation
Wang Jian Zhao Honglian Liu Bing Wang Dezhuang
Northeastern University, Computer Architecture Institute, Shenyang 110004
国际会议
2009年中国控制与决策会议(2009 Chinese Control and Decision Conference)
广西桂林
英文
653-656
2009-06-17（万方平台首次上网日期，不代表论文的发表时间）