Sample selection for visual domain adaptation via sparse coding
作者：
Highlights：
• We aim at selecting relevant source samples using sparse coding theory.
• L2,1 norm regularization is used to restrict the selection matrix.
• We do the selection process in a low dimensional shared space.
• We also exploit the target specific features.
• A unified framework is presented to learn the selection matrix and projection matrices.
摘要
Highlights•We aim at selecting relevant source samples using sparse coding theory.•L2,1 norm regularization is used to restrict the selection matrix.•We do the selection process in a low dimensional shared space.•We also exploit the target specific features.•A unified framework is presented to learn the selection matrix and projection matrices.
论文关键词：Image classification,Visual domain adaptation,Sample selection,Sparse coding,L2,1 norm
论文评审过程：Received 21 September 2015, Revised 31 March 2016, Accepted 31 March 2016, Available online 2 April 2016, Version of Record 6 April 2016.
论文官网地址：https://doi.org/10.1016/j.image.2016.03.009
