| 人脸识别中优化特征空间方法的研究 |
| |
| 引用本文： | 孙大瑞,李力杰,吴乐南. 人脸识别中优化特征空间方法的研究[J]. 信号处理, 2001, 17(6): 510-514 |
| |
| 作者姓名： | 孙大瑞 李力杰 吴乐南 |
| |
| 作者单位： | 东南大学无线电工程系,南京,210018 |
| |
| 摘 要： | 在人脸识别领域的诸多算法中,PCA(主分量分析)作为一种经典的算法,仍有其独特的生命力.本文在分析传统PCA算法的基础上,结合其它一些变形,提出一种二次PCA和NFL(最近邻特征线)级连判决的方法.针对Manchester和ORL人脸库的实验表明该算法确实提高了识别率.实验前对M锄chester人脸库进行了必要的预处理,在一定程度上弥补了光照、人脸大小、头发等的影响.
|
| 关 键 词： | 人脸识别 主分量分析 特征空间 双特征脸 最近邻特征线 联合判决 |
Face Recognition Feature Space Optimization Method |
| |
| Sun Darui,Li Lijie,Wu Lenan. Face Recognition Feature Space Optimization Method[J]. Signal Processing(China), 2001, 17(6): 510-514 |
| |
| Authors: | Sun Darui Li Lijie Wu Lenan |
| |
| Abstract: | Many algorithms have been put forward in the literature of face recognition, PCA, as a Conventional method, also has its special value. With analysis of conventional PCA, this paper integrates other algorithms based on PCA, thus puts forward a method cascading twice PCA with NFL. The experiment on Manchester and ORL face database shows the method can improve recognition rate distinctively. |
| |
| Keywords: | Face recognition PCA Feature space Dual eigenface NFL Cascading judgement |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
