改进型Gabor自商图算法及其在人脸识别中的应用

Improved Gabor Self Quotient Image Algorithm and Its Application to Face Recognition

光照变化是影响人脸识别系统性能的关键问题之一,针对该问题提出了一种改进的基于Gabor特征的自商图算法。对人脸图像采用改进的加权Gabor滤波器进行平滑的Gabor特征提取,使用自商图像的方法求取图像的光照不变特征;对得到的自商图像用直方图截断等方法进行归一化;在Extended Yale B与CMU PIE人脸库上通过基于皮尔逊相关系数的最近邻方法进行实验。实验结果表明,与传统算法相比,该算法可以大幅度提高人脸识别率。

Varying illumination remains one of the major challenges for current face recognition systems.To deal with the illumination variation problem,an improved self quotient images algorithm based on Gabor features is proposed.Firstly,the modified weighted Gabor filter is used to extract the Gabor feature of the face image,then the illumination invariant characteristics of the image is obtained by using the self quotient images algorithm.The self quotient image is normalized by histogram truncation.Finally,the experiment is carried out on Extended Yale B and CMU PIE face library by the nearest neighbor classifier based on Pearson correlation coefficient.Compared with the traditional algorithm,experimental results show that the algorithm significantly improves the face recognition rate.