Adaptive graph-based discriminative nonnegative matrix factorization for image clustering

作者：

Highlights：

• The method based on NMF is an effective dimension reduction method.

• Preserving the manifold structure of the data can improve the clustering effect.

• Adaptive graph structure can avoid the influence of noise and outliers.

• The use of partial label information can improve the clustering accuracy.

摘要

•The method based on NMF is an effective dimension reduction method.•Preserving the manifold structure of the data can improve the clustering effect.•Adaptive graph structure can avoid the influence of noise and outliers.•The use of partial label information can improve the clustering accuracy.

论文关键词：Nonnegative matrix factorization,Adaptive graph regularization,Semi-supervised learning

论文评审过程：Received 12 June 2020, Revised 28 February 2021, Accepted 23 March 2021, Available online 6 April 2021, Version of Record 10 April 2021.

论文官网地址：https://doi.org/10.1016/j.image.2021.116253