Lightweight boundary refinement module based on point supervision for semantic segmentation

作者：

Highlights：

• Creation of an effective method for segmentation without extra prior information;

• Effectively global correction of foreground edge by the novel directional field representation.

• Key point supervised learning makes BRPS lightweight from both global and local views.

• The experimental results of BRPS on four public datasets show the performance improvement.

摘要

•Creation of an effective method for segmentation without extra prior information;•Effectively global correction of foreground edge by the novel directional field representation.•Key point supervised learning makes BRPS lightweight from both global and local views.•The experimental results of BRPS on four public datasets show the performance improvement.

论文关键词：Semantic segmentation,Boundary refinement,Point supervision,Point convolution,Direction field

论文评审过程：Received 20 January 2021, Revised 19 March 2021, Accepted 4 April 2021, Available online 18 April 2021, Version of Record 21 April 2021.

论文官网地址：https://doi.org/10.1016/j.imavis.2021.104169