Large-scale and rotation-invariant template matching using adaptive radial ring code histograms

作者：

Highlights：

• We proposed a novel adaptive radial ring code histograms (ARRCH) image descriptor, which uses the radial gradient codes as the rotation-invariant feature.

• Based on ARRCH feature, we introduced a novel large-scale and rotation-invariant template matching, which utilizes the coarse-to-fine strategy to deal with large-scale change.

• Several experiments carried out on the MS COCO dataset are presented, including a parameter experiment and a large-scale and rotation change matching experiment, and some applications. The experimental results demonstrate that the proposed template matching method is more resistant to large-scale and rotation differences than most of state-of-the-art matching methods.

摘要

•We proposed a novel adaptive radial ring code histograms (ARRCH) image descriptor, which uses the radial gradient codes as the rotation-invariant feature.•Based on ARRCH feature, we introduced a novel large-scale and rotation-invariant template matching, which utilizes the coarse-to-fine strategy to deal with large-scale change.•Several experiments carried out on the MS COCO dataset are presented, including a parameter experiment and a large-scale and rotation change matching experiment, and some applications. The experimental results demonstrate that the proposed template matching method is more resistant to large-scale and rotation differences than most of state-of-the-art matching methods.

论文关键词：Template matching,Adaptive radial ring code histograms,Large-scale and rotation-invariant features

论文评审过程：Received 2 July 2015, Revised 10 May 2018, Accepted 1 March 2019, Available online 7 March 2019, Version of Record 15 March 2019.

论文官网地址：https://doi.org/10.1016/j.patcog.2019.03.003