期刊
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
卷 44, 期 11, 页码 8338-8354出版社
IEEE COMPUTER SOC
DOI: 10.1109/TPAMI.2021.3083288
关键词
Three-dimensional displays; Semantics; Memory management; Task analysis; Sampling methods; Space exploration; Feature extraction; Large-scale point clouds; semantic segmentation; random sampling; local feature aggregation
资金
- Amazon Web Services in the Oxford-Singapore Human-Machine Collaboration Programme
- China Scholarship Council (CSC) Scholarship
- Natural Science Foundation of Guangdong Province [2019A1515011271]
- Science and Technology Innovation Committee of Shenzhen Municipality [RCYX20200714114641140, JCYJ20190807152209394]
- NationalNatural Science Foundation of China [U20A20185, 61972435]
This paper addresses the problem of efficient semantic segmentation of large-scale 3D point clouds and proposes RandLA-Net, a lightweight neural architecture that relies on random point sampling and local feature aggregation for fast and accurate semantic segmentation.
We study the problem of efficient semantic segmentation of large-scale 3D point clouds. By relying on expensive sampling techniques or computationally heavy pre/post-processing steps, most existing approaches are only able to be trained and operate over small-scale point clouds. In this paper, we introduce RandLA-Net, an efficient and lightweight neural architecture to directly infer per-point semantics for large-scale point clouds. The key to our approach is to use random point sampling instead of more complex point selection approaches. Although remarkably computation and memory efficient, random sampling can discard key features by chance. To overcome this, we introduce a novel local feature aggregation module to progressively increase the receptive field for each 3D point, thereby effectively preserving geometric details. Comparative experiments show that our RandLA-Net can process 1 million points in a single pass up to 200x faster than existing approaches. Moreover, extensive experiments on five large-scale point cloud datasets, including Semantic3D, SemanticKITTI, Toronto3D, NPM3D and S3DIS, demonstrate the state-of-the-art semantic segmentation performance of our RandLA-Net.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据