Max-margin multi-scale convolutional factor analysis model with application to image classification

作者：

Highlights：

• We develop a max-margin multi-scale convolutional factor analysis model for images.

• The convolution kernels could maintain the spatial correlation among the image pixels.

• The multi-scale kernels could capture the structural information at different levels.

• Max-margin learning is explored to improve the classification performance.

• Efficient inference is performed via Gibbs method for our Bayesian model.

摘要

•We develop a max-margin multi-scale convolutional factor analysis model for images.•The convolution kernels could maintain the spatial correlation among the image pixels.•The multi-scale kernels could capture the structural information at different levels.•Max-margin learning is explored to improve the classification performance.•Efficient inference is performed via Gibbs method for our Bayesian model.

论文关键词：Factor analysis (FA),Max-margin learning,Multi-scale convolution kernels,Image classification,Synthetic aperture radar (SAR) image

论文评审过程：Received 2 February 2019, Revised 25 March 2019, Accepted 6 April 2019, Available online 26 April 2019, Version of Record 16 May 2019.

论文官网地址：https://doi.org/10.1016/j.eswa.2019.04.012