摘要
复杂的光照变化是影响人脸识别的一个关键问题,针对人脸图像中的光照变化问题,采用图像局部归一化处理方法对人脸进行光照处理,提取光照不变性特征,并进行对比实验研究.实验结果表明,局部归一化算法可以使图像保持正常光照条件下具有的局部统计特性,对消除光照影响有较好的作用.
Complicated illumination variation is still one key problem in face recognition technique. In view of illumination variation in face recognition, a model-based Local Normalization technique is adopted for the faces illumination treatment, and extracting illumination invariant feature. The contrastive experiments are done with different methods. The experiments show that the method can keep the local statistical properties of the processed images, and eliminate the effect of uneven illumination effectively and efficiently.
出处
《郑州大学学报（理学版）》
CAS
2008年第3期98-101,共4页
Journal of Zhengzhou University:Natural Science Edition
基金
东营市科技攻关项目
编号2005138-1
关键词
局部归一化处理
光照变化
人脸识别
Local Normalization technique
illumination variation
face recognition
