标题:Two rank approximations for low-rank based subspace clustering
作者:Xu, Fei ;Peng, Chong ;Hu, Yunhong ;He, Guoping
通讯作者:Peng, Chong
作者机构:[Xu, Fei ] College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao, China;[Peng, Chong ] Department of Com 更多
会议名称:10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, CISP-BMEI 2017
会议日期:October 14, 2017 - October 16, 2017
来源:Proceedings - 2017 10th International Congress on Image and Signal Processing, BioMedical Engineering and Informatics, CISP-BMEI 2017
出版年:2018
卷:2018-January
页码:1-6
DOI:10.1109/CISP-BMEI.2017.8302021
摘要:Rank approximation and minimization problem is widely applied in machine learning and computer vision. As the minimum convex envelope of the rank function, the nuclear norm is often used for rank approximation and has achieved satisfactory results in different tasks. However, the nuclear norm may not be an appropriate rank approximation especially when there are large singular values. In this paper, we propose two different functions to more accurately approximate the rank function. Then based on the low-rank representation model, we use these approximations for robust subspace clustering with desirable low rank and robustness to noise. Experimental results show the effectiveness of our proposed methods.
© 2017 IEEE.
收录类别:EI
资源类型:会议论文
TOP