Sparse multi-stage regularized feature learning for robust face recognition
作者：
Highlights：
• A new feature-extraction approach for efficient FR based on directional features.
• An augmented multi-stage or multi-regularization convex optimization.
• Results on various face databases demonstrate the effectiveness of our approach.
• The approach presents a trade-off between recognition and computational complexity.
• Three protocol experimental are tested among them the SSPP.
摘要
•A new feature-extraction approach for efficient FR based on directional features.•An augmented multi-stage or multi-regularization convex optimization.•Results on various face databases demonstrate the effectiveness of our approach.•The approach presents a trade-off between recognition and computational complexity.•Three protocol experimental are tested among them the SSPP.
论文关键词：Neural networks,Shearlets,Sparsity,Shearlet Networks,Wavelet Networks,Face recognition
论文评审过程：Available online 5 August 2014.
论文官网地址：https://doi.org/10.1016/j.eswa.2014.07.044
