基于Fisherface和SIFT特征集成的人脸识别  

Face Recognition Based on Fisherface and SIFT Features Integration

作　　者：周琳琳[1] 何中市[1] ZHOU Lin-lin;HE Zhong-shi;College of Computer Science Chongqing University;

机构地区：[1]重庆大学计算机学院,重庆400044

出　　处：《现代计算机（中旬刊）》2017年第1期58-62,共5页Modern Computer

摘　　要：针对人脸识别中人脸表情变化、光照影响、遮挡问题,提出一种基于Fisherface全局特征和SIFT局部特征集成的人脸识别方法:首先利用Fisher线性鉴别方法提取全局面部特征,然后利用SIFT算法和K-Means算法提取SIFT特征及划分子区域来构造局部面部特征,并采用概率统计的方法为子区域赋予权值来计算局部特征相似度,最后采用加权求和的方式将全局和局部面部特征并行集成来提高人脸识别的准确率。实验结果表明,该方法取得较好的识别效果,具有很好的鲁棒性。To solve the problems of facial expression changes, illumination effect and face occluded in face recognition, proposes a face recognition method based on Fisherface global and SIFT local features integration. First, uses fisher linear identification method to extract the global facial features. Then uses SIFT algorithm and K-Means clustering algorithm to extract SIFT features and divide sub domains in order to construct local facial features. In this part, method of weight calculation based on probability statistics is used to calculate local features similarity. Finally, weighted sum method is used to combine the global and local facial features in parallel manner in order to improve the accuracy of face recognition. The experimental result shows that the method achieves better recognition results and has good robustness.

关 键 词：FISHERFACE SIFT特征 K-MEANS 人脸识别 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]