Decremental Sparse Modeling Representative Selection for prototype selection
作者：
Highlights：
• A novel framework for prototype selection in subspaces is proposed.
• A recursive prototype elimination is adopted.
• It iterates a data self-representativeness coding with a block sparsity constraint.
• Classification performance after selection is assessed on five public image datasets.
• We use K-NN, Nearest subspace, Sparse Representation Classifier, and SVM.
摘要
Highlights•A novel framework for prototype selection in subspaces is proposed.•A recursive prototype elimination is adopted.•It iterates a data self-representativeness coding with a block sparsity constraint.•Classification performance after selection is assessed on five public image datasets.•We use K-NN, Nearest subspace, Sparse Representation Classifier, and SVM.
论文关键词：Prototype selection,Data self-representativeness,Subspace representation,Regularization,Block sparsity,Image classification,Object recognition
论文评审过程：Received 28 October 2014, Revised 4 May 2015, Accepted 18 May 2015, Available online 30 May 2015, Version of Record 16 July 2015.
论文官网地址：https://doi.org/10.1016/j.patcog.2015.05.018
