Discriminative sparse flexible manifold embedding with novel graph for robust visual representation and label propagation
作者：
Highlights：
• We discuss the problem of robust visual representation and label prediction.
• A new Discriminative Sparse Flexible Manifold Embedding model is proposed.
• We propose to reduce the negative effects of mixed signs on the performance.
• The proposed method is robust to outliers and noise in given data.
• We consider the similarity/dissimilarity to discover the true neighborhood.
摘要
•We discuss the problem of robust visual representation and label prediction.•A new Discriminative Sparse Flexible Manifold Embedding model is proposed.•We propose to reduce the negative effects of mixed signs on the performance.•The proposed method is robust to outliers and noise in given data.•We consider the similarity/dissimilarity to discover the true neighborhood.
论文关键词：Flexible manifold embedding,Semi-supervised learning,l2,1-Norm regularization,Novel graph construction,Robust representation and recognition
论文评审过程：Received 3 January 2016, Revised 11 July 2016, Accepted 30 July 2016, Available online 5 August 2016, Version of Record 1 September 2016.
论文官网地址：https://doi.org/10.1016/j.patcog.2016.07.042
