How to learn long-range dependencies from 3D point clouds is a challenging problem in 3D point cloud analysis. Addressing this problem, we propose a global attention network for point cloud semantic segmentation, named as GA-Net, consisting of a point-independent global attention module and a point-dependent global attention module for obtaining contextual information of 3D point clouds in this paper. The point-independent global attention module simply shares a global attention map for all 3D points. In the point-dependent global attention module, for each point, a novel random cross attention block using only two randomly sampled subsets is exploited to learn the contextual information of all the points. Additionally, we design a novel point-adaptive aggregation block to replace linear skip connection for aggregating more discriminate features. Extensive experimental results on three 3D public datasets demonstrate that our method outperforms state-of-the-art methods in most cases.

中文翻译：

---

GA-NET：点云语义分割的全局注意力网络


如何从3D点云中学习远程依赖关系是3D点云分析中的一个具有挑战性的问题。针对这个问题，本文提出了一种用于点云语义分割的全局注意力网络，称为GA-Net，由点无关全局注意力模块和点依赖全局注意力模块组成，用于获取3D点云的上下文信息。与点无关的全局注意力模块只是共享所有 3D 点的全局注意力图。在点相关的全局注意力模块中，对于每个点，利用仅使用两个随机采样子集的新颖随机交叉注意力块来学习所有点的上下文信息。此外，我们设计了一种新颖的点自适应聚合块来代替线性跳跃连接，以聚合更多区分特征。对三个 3D 公共数据集的广泛实验结果表明，我们的方法在大多数情况下优于最先进的方法。