一种基于DCT和SVM的人脸识别新方法  

Novel Face Recognition Method Based on DCT and SVM

作　　者：王孝国[1] 王景玉[2] 黄勇 张雄伟[1] 刘斌 

机构地区：[1]解放军理工大学通信工程学院电子信息工程系,江苏南京210007 [2]解放军理工大学理学院基础电子学系,江苏南京211101

出　　处：《军事通信技术》2008年第2期69-72,共4页Journal of Military Communications Technology

摘　　要：文中提出了一种新的人脸识别方法,该方法采用DCT提取人脸特征,并采用SVM对该特征进行分类识别。基于该方法,对ORL人脸库进行分类识别,仅用28个特征平均识别率达到97.4%。仿真结果表明,该方法显著地降低了特征维数和计算复杂度,明显提高了特征的可辨别能力,而且,SVM可以有效地提高分类器的分类和推广能力。A novel face recognition method was presented in this paper.In this method,Discrete Cosine Transform(DCT) was used to extract face features,and Support Vector Machine(SVM) was selected to perform face classification.Experimental results on ORL face database show that the proposed method achieves an average recognition accuracy of 97.4% using only 28 features.The computational complexity and features dimension are reduced greatly,and the dicriminant power of the features is enhanced obviously.Moreover,the classification and generalization performance of the classifier increases effectively by SVM.

关 键 词：人脸识别 离散余弦变换 支持向量机 最近邻分类器 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]