针对光照变化的人脸识别系统识别率偏低的问题,提出了基于局部模式纹理描述符的人脸识别方法。通过提取出局部模式对光照改变不敏感的纹理特性,采用距离转换度量(distanceconversionmetrics,DCM)和K均值(K-mean)算法,有效地提高了人脸识别的识别率。所提方法的有效性分别在ATR—Jaffe及Yale两大人脸库上得到了验证。实验结果表明,与其它最先进的几种方法相比,所提方法在处理非限制条件下的人脸识别问题上取得了更好的识别效果。
For problem of low recognition rate of face recognition system with illustration variation. The method Local Binary Pattern (LBP) based on local texture descriptor is proposed. Distance Conversion Metrics (DCM) and K-mean algorithm are considered as classifiers based on Local Binary Pattern (LBP). The effectiveness of the proposed method has been verified on the ATR-Jaffe and Yale face database. Experiments results show that pro- posed method has the better recognition performance under uncontrolled conditions comparing with several latest ap- proaches.