智能科学与技术学报 ›› 2021, Vol. 3 ›› Issue (3): 334-341.doi: 10.11959/j.issn.2096-6652.202134

• 专刊:目标智能检测与识别 • 上一篇    下一篇

一种基于最大相关熵和局部约束的协同表示分类器

余沁茹, 卢桂馥   

  1. 安徽工程大学计算机与信息学院,安徽 芜湖 241000
  • 修回日期:2021-07-19 出版日期:2021-09-15 发布日期:2021-09-01
  • 作者简介:余沁茹(1997− ),女,安徽工程大学计算机与信息学院硕士生,主要研究方向为图像处理
    卢桂馥(1976− ),男,博士,安徽工程大学计算机与信息学院教授,主要研究方向为计算机图形学及图像处理
  • 基金资助:
    国家自然科学基金资助项目(61976005);国家自然科学基金资助项目(61772277);安徽省自然科学基金资助项目(1908085MF183)

Collaborative representation based classifier with maximum correntropy criterion and locality constraint

Qinru YU, Guifu LU   

  1. School of Computer and Information, Anhui Polytechnic University, Wuhu 241000, China
  • Revised:2021-07-19 Online:2021-09-15 Published:2021-09-01
  • Supported by:
    The National Natural Science Foundation of China(61976005);The National Natural Science Foundation of China(61772277);The Natural Science Foundation of Anhui Province(1908085MF183)

摘要:

提出一种基于最大相关熵和局部约束的协同表示分类器(CRC/MCCLC),该分类器能同时利用最大相关熵和局部信息。一方面,通过利用最大相关熵准则,CRC/MCCLC不仅在异常值处理上比L1范数鲁棒性更高,还可以使用半二次优化技术进行更有效的计算;另一方面,CRC/MCCLC 通过使用局部信息得到近似稀疏表示,以此从训练样本中获得更多的判别信息。在 ORL、Yale 以及 AR 人脸数据集等图像数据集上的实验结果验证了CRC/MCCLC 方法的有效性。

关键词: 人脸识别, 协同表示, 稀疏表示, 最大相关熵, 局部约束

Abstract:

A method which utilizes maximum correntropy criterion and locality information called collaborative representation based classifier with maximum correntropy criterion and locality constraint (CRC/MCCLC) was proposed.On the one hand, CRC/MCCLC was not only more robust to outliers than L1 norm but also could be computed efficiently using half-quadratic optimization technique because of the use of maximum correntropy criterion.On the other hand, CRC/MCCLC could obtain more discriminative information from the training samples and could lead to an approximately sparse representation because of the use of locality information.Extensive experimental results on some image databases demonstrate that CRC/MCCLC can achieve the state-of-the-art performance on these image databases.

Key words: face recognition, collaborative representation, sparse representation, maximum correntropy criterion, locality constraint

中图分类号: 

No Suggested Reading articles found!