| 加权SRC算法在伪装人脸识别中的应用 |
| |
| 引用本文： | 甘俊英,肖娟.加权SRC算法在伪装人脸识别中的应用[J].系统工程理论与实践,2011(Z2):124-128. |
| |
| 作者姓名： | 甘俊英 肖娟 |
| |
| 作者单位： | 五邑大学信息工程学院; |
| |
| 基金项目： | 国家自然科学基金(61072127,61070167); 广东省自然科学基金(10152902001000002,S2011010001085,S2011040004211,07010869); 广东省高等学校高层次人才基金(粤教师函[2010]79号) |
| |
| 摘 要： | 针对伪装人脸识别问题,基于稀疏表示分类法(sparse representation-based classification,SRC)鲁棒性好的特性,分析了SRC算法,探讨了该算法存在的问题:稀疏解不够稀疏.提出了一种加权SRC模型,即在SRC模型基础上增加一个权值.结合Homotopy算法获得了更理想、更稀疏的解.基于AR数据库的实验结果表明,加权SRC算法的识别率随权值的变化而变化,拥有比SRC算法更高的正确识别率.
|
| 关 键 词： | 稀疏表示 伪装人脸识别 SRC 加权SRC |
Application of a weighted SRC algorithm in face recognition with disguise |
| |
| GAN Jun-ying,XIAO Juan.Application of a weighted SRC algorithm in face recognition with disguise[J].Systems Engineering —Theory & Practice,2011(Z2):124-128. |
| |
| Authors: | GAN Jun-ying XIAO Juan |
| |
| Affiliation: | GAN Jun-ying,XIAO Juan (School of Information Engineering,Wuyi University,Jiangmen 529020,China) |
| |
| Abstract: | Sparse representation-based classification(SRC) has good robustness for face recognition with disguise.At first,SRC algorithm was analyzed,and there was a problem which the solution is not sparse enough.Then,weighted SRC model was presented to add a weight value based on SRC,and more ideal and sparse solution was acquired than the original algorithm by Homotopy algorithm.Experimental results on the AR database show that recognition rate of weighted SRC is changed with the variation of weight value,and reach... |
| |
| Keywords: | sparse representation face recognition with disguise sparse representation-based classification weighted SRC |
| 本文献已被 CNKI 等数据库收录！ |
|
