摘要
提出一种基于奇异值分解和径向基函数神经网络的人脸特征提取与识别方法,来解决人脸识别中的高维、小样本问题。该方法采用奇异值分解、奇异值降维压缩、奇异值矢量标准化和奇异值矢量排序,最后得到用于识别的奇异值特征矢量。运用基于径向基函数神经网络分类器进行人脸分类识别。在ORL数据库上进行实验和数据分析表明,该方法无论是在分类的错误率上还是在学习的效率上都能表现出极好的性能。
To solve the problem of high dimension and small sample in face recognition,this paper proposed a facial feature extraction and recognition method based on singular value decomposition and radial basis function neural network.By singular value decomposition,dimension reduction and compression,vector standardization and vector sorting,we finally got the feature vectors of singular value used to identify.It used the neural network classifier based on radial basis function to classify and recognize face.The experiments and data analysis on ORL database show that,this method has good performance whether in the error rate of classification or in the learning efficiency.
出处
《计算机科学》
CSCD
北大核心
2012年第B06期566-569,共4页
Computer Science
基金
广东省自然科学基金(S2011020002719
10152800001000016)资助
关键词
RBF神经网络
奇异值分解
特征提取
人脸识别
Radial basis function neural network; Singular value decomposition; Feature extraction; Face recognition