A generalised framework for saliency-based point feature detection

作者：

Highlights：

• A novel multi-modal histogram-based salient point detector is proposed.

• The general formulation allows salient point detection in both images and 3D data.

• The derivative-based approach improves upon existing intensity based 2D approaches.

• The 3D detector naturally operates on both the geometry and texture of 3D scenes.

• Increased 2D-2D and 2D-3D repeatability demonstrated on indoor and outdoor scenes.

摘要

•A novel multi-modal histogram-based salient point detector is proposed.•The general formulation allows salient point detection in both images and 3D data.•The derivative-based approach improves upon existing intensity based 2D approaches.•The 3D detector naturally operates on both the geometry and texture of 3D scenes.•Increased 2D-2D and 2D-3D repeatability demonstrated on indoor and outdoor scenes.

论文关键词：Point detection,Feature detection,Feature matching,2D-3D registration,Saliency

论文评审过程：Received 30 November 2015, Revised 26 August 2016, Accepted 13 September 2016, Available online 14 September 2016, Version of Record 18 March 2017.

论文官网地址：https://doi.org/10.1016/j.cviu.2016.09.008