摘要
提出了一种基于小波分解和分类的人脸识别算法;算法首先对训练样本进行小波分解,以方差最大之小波系数间相关系数作为分类距离,对样本进行分类,并确定每类图像的类心;人脸识别过程首先寻找与测试样本匹配程度最高的类心图像,然后在该类心图像所在类中寻找最佳匹配图像,从而减少存储空间和计算时间,而且分类和确定类心均是离线操作,从而该算法显著加快了人脸识别速度;实验结果表明,算法有效。
An efficient method which based on the face classification is being used to the face recognition. The train images are being transformed by the Daubechies Wavelet and find the max variance of them. The cross correlation among the wavelet is being used as the dis- tance to cluster, and the central face image of per classification is found. The face recognition method： first of all, finding the central face image which has the best matching in all the central face images; then finding the perfect matching face image from the classification which has the central face image. The face classification and the finding for the central image are the off--line works, so they accelerate the face recognition. The experiment indicates that the method has advantage over other methods.
出处
《计算机测量与控制》
CSCD
北大核心
2009年第1期167-169,共3页
Computer Measurement &Control
基金
国家自然科学基金资助项目(70471065)
上海市重点学科建设资助项目(T0502)
上海市自然科学基金资助项目(06ZR14144)
关键词
人脸识别
人脸分类
聚类分析
相关系数
小波
face classification
face recognition
clustering
correlation
wavelet
