| 基于谱回归判别分析的LPP算法 |
| |
| 引用本文： | 杨凡,张银玲,牛静.基于谱回归判别分析的LPP算法[J].微型机与应用,2012(16):38-41. |
| |
| 作者姓名： | 杨凡 张银玲 牛静 |
| |
| 作者单位： | 浙江师范大学数理与信息工程学院 |
| |
| 基金项目： | 浙江师范大学计算机软件与理论省级重中之重学科开放基金;浙江省大学生科技创新活动计划(新苗人才计划)项目编号:2011R404054 |
| |
| 摘 要： | 判别局部保持投影DLPP算法在计算过程中需要解决稠密矩阵特征分解问题,这使得该算法在时间和内存上消耗都非常高。谱回归判别分析SRDA算法可以有效的节省时间和内存的消耗。基于SRDA,提出一种改进的局部保持投影LPP算法——谱回归判别局部保持投影算法SRDLPP。实验结果表明,该算法可以提高识别率,同时降低时间和内存消耗。
|
| 关 键 词： | 判别局部保持投影 局部保持投影算法 谱回归判别分析 人脸识别 |
LPP algorithm based on spectral regression discriminant analysis |
| |
| Yang Fan, Zhang Yinling, Niu Jing.LPP algorithm based on spectral regression discriminant analysis[J].Microcomputer & its Applications,2012(16):38-41. |
| |
| Authors: | Yang Fan Zhang Yinling Niu Jing |
| |
| Affiliation: | (College of Mathematics Physics and Information Engineering of Zhejiang Normal University, Jinhua 321004, China) |
| |
| Abstract: | The computation of Discriminant Locality Preserving Projection (DLPP) algorithm involves dense matrices eigen-decomposition which can be computationally expensive both in time and memory. Spectral Regression Discriminant Analysis (SRDA) algorithm can save both time and memory efficiently. Inspired by SRDA, we propose a novel improvement algorithm for LPP: called Spectral Regression Discriminant Locality Preserving Projection (SRDLPP). Experimental results on face recognition demonstrate that it can improve recognition rate and meanwhile save time and memory. |
| |
| Keywords: | locality preserving projection discriminant locality preserving projection spectral regression discriminant analysis face recognition |
| 本文献已被 CNKI 等数据库收录！ |
