| 偏最小二乘回归分析与特征融合在图像识别中的应用 |
| |
| 引用本文： | 杨茂龙,孙权森,夏德深.偏最小二乘回归分析与特征融合在图像识别中的应用[J].计算机工程与应用,2009,45(26):187-189. |
| |
| 作者姓名： | 杨茂龙 孙权森 夏德深 |
| |
| 作者单位： | 1. 南京国际关系学院六系,南京,210031;南京理工大学计算机学院,南京,210094
2. 南京理工大学计算机学院,南京,210094 |
| |
| 摘 要： | 为了更有效地进行图像识别,对同一模式的不同特征进行融合是有效途径。讨论了偏最小二乘法及其改进算法、特征融合方法在图像识别中的应用。首先讨论了偏最小二乘法的基本原理和非迭代偏最小二乘法、基于共轭正交的偏最小二乘法用于特征抽取的原理和特点,给出了三种特征融合方法,在ORL与Yale人脸库上的实验结果表明进行对用PLS抽取的特征融合后可以有效地进行图像识别。
|
| 关 键 词： | 偏最小二乘 特征融合 图像识别 |
| 收稿时间： | 2008-5-15 |
| 修稿时间： | 2008-7-28
|
Image recognition based on partial least squares regression and feature fusion |
| |
| YANG Mao-long,SUN Quan-sen,XIA De-shen.Image recognition based on partial least squares regression and feature fusion[J].Computer Engineering and Applications,2009,45(26):187-189. |
| |
| Authors: | YANG Mao-long SUN Quan-sen XIA De-shen |
| |
| Affiliation: | YANG Mao-long1,2,SUN Quan-sen2,XIA De-shen2 1.International Studies University,Nanjing 210031,China 2.Department of Computer Science,Nanjing University of Science & Technology,Nanjing 210094,China |
| |
| Abstract: | It is an effective approach for image recognition more efficiently by fusing different features of a pattern.The Partial Least Squares regression(PLS) method and its improved algorithms are discussed,as well as feature fusion methods and their application.The theories of feature extracting by PLS,non-iterative PLS and orthonormalized PLS are discussed,and the three feature fusion methods are given,too.The experiment results on ORL and Yale face image database have shown that the fused feature can achieve go... |
| |
| Keywords: | Partial Least Squares（PLS） future fusion image recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
| 点击此处可从《计算机工程与应用》浏览原始摘要信息 |
|
点击此处可从《计算机工程与应用》下载全文 |
|
