Transformer architecture based on the attention mechanism achieves impressive results in natural language processing (NLP) tasks. This paper transfers the successful experience to a 3D point cloud segmentation task. Inspired by newly proposed 3D Transformer neural networks, this paper introduces a new Transformer‐based module, which is called Local Geo‐Transformer. To alleviate the heavy memory consumption of the original Transformer, Local Geo‐Transformer only performs the attention mechanism in local regions. It is designed to mitigate the information loss caused by the subsampling of point clouds for segmentation. Global Geo‐Transformer is proposed to exploit the global relationships between features with the lowest resolution. The new architecture is validated on a masonry bridge dataset developed by the authors for their earlier work on a previous segmentation network called BridgeNet. The new version of the network with Transformer architecture, BridgeNetv2, outperforms BridgeNet in all metrics. BridgeNetv2 is also shown to be lightweight and memory efficient, well‐adapted to large‐scale point cloud segmentation tasks in civil engineering.

中文翻译：

---

一种基于 Transformer 的轻量级神经网络，用于大规模砖石拱桥点云分割

基于注意力机制的 Transformer 架构在自然语言处理（NLP）任务中取得了令人印象深刻的结果。本文将成功的经验转移到3D点云分割任务中。受新提出的 3D Transformer 神经网络的启发，本文引入了一种新的基于 Transformer 的模块，称为 Local Geo-Transformer。为了减轻原始 Transformer 的大量内存消耗，Local Geo-Transformer 仅在局部区域执行注意力机制。它旨在减轻点云二次采样以进行分割所造成的信息丢失。 Global Geo-Transformer 的提出是为了利用最低分辨率的特征之间的全局关系。新的架构在砖石桥梁数据集上进行了验证，该数据集是作者为之前在名为 BridgeNet 的分段网络上所做的早期工作而开发的。采用 Transformer 架构的新版本网络 BridgeNetv2 在所有指标上都优于 BridgeNet。 BridgeNetv2 还被证明是轻量级且内存高效的，非常适合土木工程中的大规模点云分割任务。