Fusion of complex networks and randomized neural networks for texture analysis

作者：

Highlights：

• A high texture descriptor is proposed.

• It is based on the fusion of complex networks and randomized neural networks.

• Texture images are modelled as complex networks with different scales.

• Network topological characteristics are used to train a randomized neural network.

• The output weights of the randomized neural network are used as a texture descriptors.

摘要

•A high texture descriptor is proposed.•It is based on the fusion of complex networks and randomized neural networks.•Texture images are modelled as complex networks with different scales.•Network topological characteristics are used to train a randomized neural network.•The output weights of the randomized neural network are used as a texture descriptors.

论文关键词：Randomized neural networks,Complex networks,Texture analysis,Feature extraction

论文评审过程：Received 31 January 2019, Revised 13 November 2019, Accepted 24 December 2019, Available online 2 January 2020, Version of Record 18 February 2020.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.107189