Constructing Deep Sparse Coding Network for image classification

作者：

Highlights：

• We propose a novel deep model combining the advantages of CNN and sprase-coding technique.

• Image representation can be obtained directly from raw pixels.

• Multi-scale and Multi-Locality extensions are proposed to boost the recognition accuracy.

• The classification accuracy is very promising among traditional unsupervised methods.

摘要

Highlights•We propose a novel deep model combining the advantages of CNN and sprase-coding technique.•Image representation can be obtained directly from raw pixels.•Multi-scale and Multi-Locality extensions are proposed to boost the recognition accuracy.•The classification accuracy is very promising among traditional unsupervised methods.

论文关键词：Sparse Coding,Deep Model,Multi-scale,Multi-locality,Image classification

论文评审过程：Received 27 July 2015, Revised 25 October 2016, Accepted 27 October 2016, Available online 5 November 2016, Version of Record 16 November 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.10.032