| 基于PCA和LDA结合的人脸识别方法的比较研究 |
| |
| 引用本文： | 粟秀尹,曾庆宁,王晓燕.基于PCA和LDA结合的人脸识别方法的比较研究[J].计算机工程与设计,2012,33(9):3574-3578. |
| |
| 作者姓名： | 粟秀尹 曾庆宁 王晓燕 |
| |
| 作者单位： | 桂林电子科技大学信息与通信学院电子信息工程系,广西桂林,541004 |
| |
| 基金项目： | 广西自然科学基金项目(A053232);桂林电子科技大学人才专项基金项目(UF11012Y);广西区研究生创新基金项目(2011105950810M16) |
| |
| 摘 要： | 针对用于人脸识别的主分量分析(PCA)计算量大、识别效果不佳和分类时间长,线性判别分析(LDA)存在小样本问题,比较研究几种基于PCA和LDA的人脸识别方法,这些方法包括PCA+ LDA、2DPCA+ DLDA和2DPCA+2DLDA.在理论和实验上比较研究这些方法,且在ORL和Yale人脸库的实验数据表明,结合后的方法比结合前的方法分类效果好,分类时间短,且在这几种方法中,2DPCA+2DLDA的效果最佳.
|
| 关 键 词： | 主分量分析 线性判别分析 分类效果 分类时间 |
Several combination methods of face recognition based on PCA and LDA |
| |
| SU Xiu-yin
,
ZENG Qing-ning
,
WANG Xiao-yan.Several combination methods of face recognition based on PCA and LDA[J].Computer Engineering and Design,2012,33(9):3574-3578. |
| |
| Authors: | SU Xiu-yin ZENG Qing-ning WANG Xiao-yan |
| |
| Affiliation: | (Department of Electronic Information Engineering,College of Information and Communications, Guilin University of Electronic Technology,Guilin 541004,China) |
| |
| Abstract: | To solve principal component analysis(PCA) for face recognition of large calculation,poorly identified and classified for a long time,linear discriminate analysis(LDA) of a small sample size problem,several combination methods of face recognition based on PCA and LDA are studied.These methods include PCA+LDA,2DPCA+DLDA and 2DPCA+2DLDA.These methods are compared in the theory and experiment,and the experimental data in the ORL and Yale face database show that the combined method are better than the former form on the effect of classification,the short time of classification.In these methods,the 2DPCA +2DLDA method is the best on classification. |
| |
| Keywords: | principal component analysis linear discriminate analysis classification results classification time |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
