【作者】 陈松； 蔡建立；

【Author】 CHEN Song, CAI Jian-li (School of Information Science and Technology, Xiamen University, Xiamen 361005, China)

【机构】 厦门大学信息科学与技术学院；

【摘要】 该文提出了基于小波特征以及核主元分析(KPCA)的特点获取多组人脸特征,再设计线性支持向量机(SVM)分类器进行分类识别。首先利用小波变换将人脸图像分解成不同的频率子带, 对其中的低频平滑子带运用KPCA求取多组特征向量,将获得的特征向量划分为主特征向量和次特征向量,再由主特征向量和次特征向量组合成最终的分类特征向量。为了获得最优的识别效果,在ORL标准人脸库上进行相应的试验。试验结果表明:该方法可以取得较好的识别结果。更多还原

【Abstract】 Feature selection is very important for face recognition. In this paper, an efficient method based on Wavelet Feature and kernel principle component analysis (KPCA) is proposed for face feature selection and a linear support vector machine (SVM) classifier is designed for face recognition. Firstly, wavelet transform is used to preprocess the original training samples and three groups of wavelet features which correspond to low frequency are obtained. Then KPCA is performed in wavelet features to obtain many classes of feature vectors. After that, they are divided into master feature vectors and secondary feature vectors. Finally, the classified feature vectors were combined by master feature vectors and secondary feature vectors. Experimental results on ORL face databases indicate the effectiveness of the proposed method.更多还原