| 基于核独立成分分析的人脸识别 |
| |
| 引用本文： | 张燕昆,刘重庆.基于核独立成分分析的人脸识别[J].光学技术,2004,30(5):613-615. |
| |
| 作者姓名： | 张燕昆 刘重庆 |
| |
| 作者单位： | 上海交通大学,图像处理与模式识别研究所,上海,200030 |
| |
| 摘 要： | 研究一种基于核独立成分分析的人脸识别方法。利用支持向量机的核函数思想,将原始人脸图像向量映射到高维特征空间,然后在高维特征空间中进行独立成分分析(ICA),提取非线性独立成分作为特征向量进行分类识别。实验结果表明该方法要比常规的基于ICA和PCA的人脸识别算法的识别率要高。
|
| 关 键 词： | 人脸识别 主元分析 独立成分分析 核主元分析 |
| 文章编号： | 1002-1582(2004)05-0613-03 |
| 修稿时间： | 2003年10月20 |
Face recognition based on kernel independent components analysis |
| |
| ZHANG Yan-kun,LIU Chong-qing.Face recognition based on kernel independent components analysis[J].Optical Technique,2004,30(5):613-615. |
| |
| Authors: | ZHANG Yan-kun LIU Chong-qing |
| |
| Abstract: | A face recognition methods based on kernel independent component analysis(KICA) is proposed. Adopted the kernel function used in support vector machine, the origin face image vector is mapped into high dimensional feature space, then do the independent component analysis in high dimensional feature space to extract nonlinear independent features. The result shows that the proposed method obtained high recognition rate than that of the traditional ICA and PCA methods. |
| |
| Keywords: | face recognition principal component analysis independent component analysis kernel principal component analysis |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
