Paying attention for adjacent areas: Learning discriminative features for large-scale 3D scene segmentation

作者：

Highlights：

• We propose an ARmodule which contains two novel attention blocks:large-scale support spatial attention the extended channel attention. (N∼105)..

• We propose two loss functions to solve the intra-class inconsistency and inter-class indistinction for long-tailed distribution of 3D scenes.

• The proposed ARNet integrates AR module and loss functions in an end-to-end manner, which achieves the SOTA results on indoor and outdoor datasets.

摘要

•We propose an ARmodule which contains two novel attention blocks:large-scale support spatial attention the extended channel attention. (N∼105)..•We propose two loss functions to solve the intra-class inconsistency and inter-class indistinction for long-tailed distribution of 3D scenes.•The proposed ARNet integrates AR module and loss functions in an end-to-end manner, which achieves the SOTA results on indoor and outdoor datasets.

论文关键词：Large-scale 3D point clouds,Attention,Long-tailed distribution,Segmentation

论文评审过程：Received 25 March 2021, Revised 7 April 2022, Accepted 19 April 2022, Available online 20 April 2022, Version of Record 28 April 2022.

论文官网地址：https://doi.org/10.1016/j.patcog.2022.108722