A pooled Object Bank descriptor for image scene classification

作者：

Highlights：

• We propose a threshold value filter pooling method.

• The pooling is capable of avoiding the accumulation of noise in histogram-pooling.

• We propose a Matthew effect normalization method to highlight the useful information.

• We apply such two methods in a dimensionality reduction framework.

• The framework can reduce dimensionality and increase classification accuracy.

摘要

•We propose a threshold value filter pooling method.•The pooling is capable of avoiding the accumulation of noise in histogram-pooling.•We propose a Matthew effect normalization method to highlight the useful information.•We apply such two methods in a dimensionality reduction framework.•The framework can reduce dimensionality and increase classification accuracy.

论文关键词：Image classification,Object Bank,Dimensionality reduction,Pooling,Image feature

论文评审过程：Received 12 January 2017, Revised 12 October 2017, Accepted 30 October 2017, Available online 31 October 2017, Version of Record 13 November 2017.

论文官网地址：https://doi.org/10.1016/j.eswa.2017.10.057