Method based on wavelet multiresolution analysis and KPCA for face recognition
基于小波分析与KPCA的人脸识别方法

Feature selection is very important for face recognition.The valuable facial low-frequency features can be obtained by wavelet multiresolution analysis and the non-linear features kernel principle component analysis(KPCA) can be extracted from initial face images.In this paper,an efficient method based on wavelet multiresolution analysis and KPCA was proposed for face feature selection and a linear support vector machine(SVM) classifier was designed for face recognition.Experimental results on UMIST face databases indicate the effectiveness of the proposed method.