Evaluation of ground distances and features in EMD-based GMM matching for texture classification

作者：

Highlights：

• We present a comprehensive study of ground distances and image features in EMD-based GMM Matching for texture classification.

• An improved Gaussian embedding distance is proposed to compare Gaussians.

• The experimental results show that our method can achieve state-of-the-art performance.

摘要

Highlights•We present a comprehensive study of ground distances and image features in EMD-based GMM Matching for texture classification.•An improved Gaussian embedding distance is proposed to compare Gaussians.•The experimental results show that our method can achieve state-of-the-art performance.

论文关键词：Texture classification,Earth Mover׳s Distance,Gaussian mixture models,Ground distances,Image features

论文评审过程：Received 11 September 2015, Revised 29 December 2015, Accepted 2 March 2016, Available online 17 March 2016, Version of Record 6 May 2016.

论文官网地址：https://doi.org/10.1016/j.patcog.2016.03.001