Sparse feature selection based on graph Laplacian for web image annotation

作者：

Highlights：

• Spare feature selection method based on l2,1/2-matix norm is proposed.

• Graph Laplacian based semi-supervised learning is exploited.

• A effective algorithm for optimizing the objective function is introduced.

• The convergence of the algorithm is proven.

• Experiments demonstrate that the method is suitable for web image annotation.

摘要

•Spare feature selection method based on l2,1/2-matix norm is proposed.•Graph Laplacian based semi-supervised learning is exploited.•A effective algorithm for optimizing the objective function is introduced.•The convergence of the algorithm is proven.•Experiments demonstrate that the method is suitable for web image annotation.

论文关键词：Web image annotation,Sparse feature selection,l2,1/2-matrix norm,Semi-supervised learning,Graph Laplacian

论文评审过程：Received 31 July 2013, Revised 6 December 2013, Accepted 30 December 2013, Available online 18 January 2014.

论文官网地址：https://doi.org/10.1016/j.imavis.2013.12.013