摘要
1 Introduction and main contributions Recently,low-rank representation(LRR)has been introduced and applied to capture the underlying low-dimensional struc-ture of data in pattern recognition and computer vision.LRR could detect the corruptions automatically during data recov-ery under mild conditions[1].In the sense of simultaneously optimizing the data representation,LRR is claimed to well capture the global structure of the data[2].
基金
This work was supported by the Basic Science Research of Shaanxi Province(2018JQ1038,2018JQ5059)
the Special Fund for Basic Scientific Research of Central Colleagues,Chang’an University(310812171006).