{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,2]],"date-time":"2026-04-02T16:29:49Z","timestamp":1775147389234,"version":"3.50.1"},"reference-count":36,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T00:00:00Z","timestamp":1762387200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Comput. Sci."],"abstract":"<jats:p>Point cloud semantic segmentation technology for road scenes plays an important role in the field of autonomous driving. However, accurate semantic segmentation of large-scale and non-uniformly dense LiDAR road point clouds still faces severe challenges. To this end, this paper proposes a road point cloud semantic segmentation algorithm called RWAFormer. First, a sparse tensor feature encoding module (STFE) is introduced to enhance the network\u2019s ability to extract local features of point clouds. Secondly, a radial window attention module (RWA) is designed to dynamically select the neighborhood window size according to the distance of the point cloud data from the center point, effectively aggregating the information of long-distance sparse point clouds to the adjacent dense areas, significantly improving the segmentation effect of long-distance point clouds. Experimental results show that our method achieves an average intersection over union (mIoU) of 75.3 and 82.0% on the Semantickitti and Nuscenes datasets, and an accuracy (Acc) of 94.5 and 97.4%. These results validate the effectiveness and superiority of RWAFormer in road point cloud semantic segmentation.<\/jats:p>","DOI":"10.3389\/fcomp.2025.1542813","type":"journal-article","created":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T11:37:53Z","timestamp":1762429073000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":1,"title":["RWAFormer: a lightweight road LiDAR point cloud segmentation network based on transformer"],"prefix":"10.3389","volume":"7","author":[{"given":"Zirui","family":"Li","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lei","family":"Chen","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Ying","family":"Liu","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shuang","family":"Zhao","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Qinghe","family":"Guan","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1965","published-online":{"date-parts":[[2025,11,6]]},"reference":[{"key":"ref1","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV.2019.00939","article-title":"SemanticKITTI: a dataset for semantic scene understanding of LiDAR sequences","author":"Behley","year":"2019"},{"key":"ref2","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR42600.2020.01164","article-title":"nuscenes: a multimodal dataset for autonomous driving","author":"Caesar","year":"2020"},{"key":"ref3","author":"Cen","year":"2022"},{"key":"ref4","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2017.195","article-title":"Xception: deep learning with depthwise separable convolutions","author":"Chollet","year":"2017"},{"key":"ref5","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2019.00319","article-title":"4D spatio-temporal convnets: Minkowski convolutional neural networks","author":"Choy","year":"2019"},{"key":"ref6","author":"Feng","year":"2024"},{"key":"ref7","first-page":"3354","article-title":"Are we ready for autonomous driving? The Kitti vision benchmark suite","author":"Geiger","year":"2012"},{"key":"ref8","first-page":"11108","article-title":"Randla-net: efficient semantic segmentation of large-scale point clouds","author":"Hu","year":"2020"},{"key":"ref9","doi-asserted-by":"crossref","first-page":"6844","DOI":"10.1007\/s10489-022-03930-5","article-title":"Semantic segmentation of 3D LiDAR data using deep learning: a review of projection-based methods","volume":"53","author":"Jhaldiyal","year":"2023","journal-title":"Appl. Intell."},{"key":"ref10","author":"Jia","year":"2021"},{"key":"ref11","first-page":"11096","article-title":"Lidarmultinet: towards a unified multi-task network for lidar perception","volume":"7","author":"Li","year":"2022","journal-title":"IEEE Robot. Autom. Lett."},{"key":"ref12","first-page":"8895","article-title":"Relation-shape convolutional neural network for point cloud analysis","author":"Liu","year":"2019"},{"key":"ref13","first-page":"10012","article-title":"Swin transformer: hierarchical vision transformer using shifted windows","author":"Liu","year":"2021"},{"key":"ref14","author":"Loshchilov","year":"2017"},{"key":"ref15","year":"2024"},{"key":"ref16","first-page":"336","article-title":"Point cloud map generation and localization for autonomous vehicles using 3D LiDAR scans","author":"Poulose","year":"2022"},{"key":"ref17","first-page":"652","article-title":"Pointnet: deep learning on point sets for 3D classification and segmentation","author":"Qi","year":"2017"},{"key":"ref18","first-page":"5105","article-title":"Pointnet++: deep hierarchical feature learning on point sets in a metric space","volume":"30","author":"Qi","year":"2017","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref19","author":"Rao","year":"2022"},{"key":"ref20","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-24574-4_28","article-title":"U-net: convolutional networks for biomedical image segmentation","author":"Ronneberger","year":"2015"},{"key":"ref21","year":"2024"},{"key":"ref22","volume-title":"Research on registration methods for 3D point cloud data","author":"Song","year":"2014"},{"key":"ref23","first-page":"685","article-title":"Searching efficient 3D architectures with sparse point-voxel convolution","author":"Tang","year":"2020"},{"key":"ref24","first-page":"6411","article-title":"Kpconv: flexible and deformable convolution for point clouds","author":"Thomas","year":"2019"},{"key":"ref25","first-page":"10347","article-title":"Training data-efficient image transformers & distillation through attention","author":"Touvron","year":"2021"},{"key":"ref26","first-page":"1","article-title":"DALES: a large-scale aerial LiDAR data set for semantic segmentation","author":"Varney","year":"2020"},{"key":"ref27","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1145\/3326362","article-title":"Dynamic graph CNN for learning on point clouds","volume":"38","author":"Wang","year":"2019","journal-title":"ACM Trans. Graph."},{"key":"ref28","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/TIM.2023.3292948","article-title":"Real-time semantic segmentation of LiDAR point clouds on edge devices for unmanned systems","volume":"72","author":"Wang","year":"2023","journal-title":"IEEE Trans. Instrum. Meas."},{"key":"ref29","author":"Wu","year":"2024"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"112945","DOI":"10.1016\/j.scienta.2024.112945","article-title":"Individual nursery trees classification and segmentation using a point cloud-based neural network with dense connection pattern","volume":"328","author":"Xu","year":"2024","journal-title":"Sci. Hortic."},{"key":"ref31","first-page":"107567","article-title":"Rotation invariant point cloud classification: where local geometry meets global topology","volume":"109","author":"Zhang","year":"2021","journal-title":"Pattern Recogn."},{"key":"ref32","first-page":"28","article-title":"A review of deep learning-based point cloud semantic segmentation","volume":"57","author":"Zhang","year":"2020","journal-title":"Prog. Laser Optoelectron."},{"key":"ref33","doi-asserted-by":"publisher","first-page":"4471","DOI":"10.3390\/rs14184471","article-title":"SVASeg: sparse voxel-based attention for 3D LiDAR point cloud semantic segmentation","volume":"14","author":"Zhao","year":"2022","journal-title":"Remote Sens"},{"key":"ref34","first-page":"4490","article-title":"Voxelnet: end-to-end learning for point cloud based 3d object detection","author":"Zhou","year":"2018"},{"key":"ref35","first-page":"9527","article-title":"Cylindrical and asymmetrical 3D convolution networks for LiDAR-based perception","volume":"44","author":"Zhu","year":"2021","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref36","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR46437.2021.00981","article-title":"Cylindrical and asymmetrical 3D convolution networks for LiDAR segmentation","author":"Zhu","year":"2021"}],"container-title":["Frontiers in Computer Science"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2025.1542813\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,11,6]],"date-time":"2025-11-06T11:37:56Z","timestamp":1762429076000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fcomp.2025.1542813\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,6]]},"references-count":36,"alternative-id":["10.3389\/fcomp.2025.1542813"],"URL":"https:\/\/doi.org\/10.3389\/fcomp.2025.1542813","relation":{},"ISSN":["2624-9898"],"issn-type":[{"value":"2624-9898","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,11,6]]},"article-number":"1542813"}}