摘要
介绍了在Radon变换下的图像矩特征的抽取方法,并得到图像的矩特征矩阵;进而对矩特征矩阵按行向量进行小波变换组成矩-小波描述子特征矩阵,采用矩阵的加权欧氏距离作为人脸图像的匹配识别的算法,产生较好的结果。
The writers put forward a method of extracting images moment-feature which can get a moment ?feature matrix of image under Radon transform. Then it applies Wavelet transform to its row vector and get moment-wavelet descriptors feature matrix, and uses the weighted Euclidean-distance as the method of recognizing faces image. At last this article gives a result by the experiments of using this matrix shows that the method has the character of high precision.
出处
《上海师范大学学报(自然科学版)》
2004年第2期42-46,共5页
Journal of Shanghai Normal University(Natural Sciences)
基金
上海市教委"人脸图像识别研究"(CL200105)