Exploring space–frequency co-occurrences via local quantized patterns for texture representation

作者：

Highlights：

• Two-channel space–frequency feature spaces are designed to obtain invariant features.

• Two types of quantization via global thresholding are developed.

• The joint space–frequency coding is explored to capture a richer description.

• Our method outperforms the state-of-the-art on three well-known texture datasets.

摘要

Highlights•Two-channel space–frequency feature spaces are designed to obtain invariant features.•Two types of quantization via global thresholding are developed.•The joint space–frequency coding is explored to capture a richer description.•Our method outperforms the state-of-the-art on three well-known texture datasets.

论文关键词：Texture classification,Image descriptor,Local Binary Pattern (LBP),Textons,Local Fourier Transform (FT)

论文评审过程：Received 5 July 2013, Revised 10 October 2014, Accepted 5 March 2015, Available online 13 March 2015.

论文官网地址：https://doi.org/10.1016/j.patcog.2015.03.003