| 一种基于核函数特征提取改进方法的应用 |
| |
| 引用本文： | 李德启,刘传领. 一种基于核函数特征提取改进方法的应用[J]. 计算机应用研究, 2011, 28(8): 3185-3187. DOI: 10.3969/j.issn.1001-3695.2011.08.107 |
| |
| 作者姓名： | 李德启 刘传领 |
| |
| 作者单位： | 1. 商丘职业技术学院,河南商丘,476000
2. 南京理工大学计算机学院,南京,210094 |
| |
| 摘 要： | 针对标准的核典型相关分析（KCCA）方法在对训练样本增大的情况下相应计算机复杂度剧增、内存占用量大的缺陷，在对标准的KCCA特征提取方法分析推导的基础上，提出了一种改进的核函数特征提取方法。该方法首先根据特征值的大小对训练样本重要程度进行判断，进而完成对应特征向量的提取;然后通过与SVDD分类器的结合，在对图像识别率影响不大的情况下，提升了对图像特征提取的效率,节省了系统的存储量；最后通过在Yale标准人脸库上进行仿真对比实验，验证了该方法的可行性，从而为提高图像模式识别效率提供了一种有效的途径。
|
| 关 键 词： | 核典型相关分析 特征提取 计算复杂度 内存占用量 识别率 |
| 收稿时间： | 2011-01-20 |
| 修稿时间： | 2011-07-02 |
Feature extraction based on kernel function and application of improved methods |
| |
| LI De-qi,LIU Chuan-ling. Feature extraction based on kernel function and application of improved methods[J]. Application Research of Computers, 2011, 28(8): 3185-3187. DOI: 10.3969/j.issn.1001-3695.2011.08.107 |
| |
| Authors: | LI De-qi LIU Chuan-ling |
| |
| Affiliation: | LI De-qi1,LIU Chuan-ling2(1.Shangqiu Polytechnic,Shangqiu Henan 476000,China,2.College of Computer Science,Nanjing University of Technology,Nanjing 210094,China) |
| |
| Abstract: | For the standard KCCA method in the case of training samples increases the complexity of the corresponding surge in the computer memory occupied by a large quantity of defects,based on the standard derivation of KCCA feature extraction methods,this paper proposed an improved feature extraction method of nuclear function.In this method,the value based on characteristics of the training sample size to judge the degree of importance,and then completed the corresponding eigenvectors,and combined with the SVDD c... |
| |
| Keywords: | kernel canonical correlation analysis(KCCA) feature extraction computational complexity memory footprint recognition rate |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《计算机应用研究》浏览原始摘要信息 |
|
点击此处可从《计算机应用研究》下载全文 |
|
