面向人脸识别的分块自适应光照处理算法

Adaptive Illumination Algorithm Based on Blocks for Face Image Recognition

为改善复杂光照条件下的人脸识别率急剧下降问题,提出一种基于双标准图的分块自适应Gamma算法(DRHGIC)。首先根据人脸光照分布特点,将人脸进行分块,然后对不同的块采用基于双标准图的方法搜索最优Gamma值,最后对人脸图片进行自适应Gamma变换。实验证明,该算法能够有效地提高复杂光照条件下的人脸识别率。

In order to improve the rate of face recognition in complex illumination condition, it proposes an adaptive Gamma algorithm based on double standard pictures and blocks （DRHGIC）. According to the illumination distribution of face, images are divided into different blocks. Then to every blocks, an optimal Gamma value based on double standard images is used to implement adaptive Gamma transform. Experiments show that this algorithm can effectively improve the rate of face recognition under complicated illumination conditions.