BRoPH: An efficient and compact binary descriptor for 3D point clouds

作者：

Highlights：

• A repeatable and stable Local Reference Frame is presented for descriptor generation.

• A novel binary local descriptor called BRoPH is proposed for 3D point clouds.

• Extensive experiments on four datasets with various data qualities are carried out.

• BRoPH gets the best efficient and compact results against selected floating methods.

摘要

•A repeatable and stable Local Reference Frame is presented for descriptor generation.•A novel binary local descriptor called BRoPH is proposed for 3D point clouds.•Extensive experiments on four datasets with various data qualities are carried out.•BRoPH gets the best efficient and compact results against selected floating methods.

论文关键词：Local reference frame,Binary feature descriptor,Object recognition,Multi-view&multi-scale

论文评审过程：Received 9 March 2017, Revised 20 November 2017, Accepted 30 November 2017, Available online 2 December 2017, Version of Record 21 December 2017.

论文官网地址：https://doi.org/10.1016/j.patcog.2017.11.029