基于SVM的人脸识别系统的研究

Face Recognition Based on Support Vector Machines

该文使用PCA来提取人脸的特征,克服了K-L算法计算量大,计算耗时长的缺点。实验表明该算法减少了特征提取时间。然后运用SVM来进行人脸识别。将两类分类问题的识别策略应用到多类分类问题。

This paper extracts face feature from face picture based on PCA algorithm and overcomes the shortage that it takes long time to calculate the result based on K-L algorithm. The experiment result shows that much time has been saved by using PCA algorithm in featue extracting. In the face recognition based on SVM the recognition strategy for two-class classification is developed to multi-class classificaiton. The recognition experiment result show a good performance in ORE face database.