PointGS: Semantic-Consistent Unsupervised 3D Point Cloud Segmentation with 3D Gaussian Splatting
PointGS: 基于3D高斯散射的语义一致无监督3D点云分割
机构 * Key Laboratory of Big Data & Artificial Intelligence in Transportation (Beijing Jiaotong University), Ministry of Education(大数据与人工智能在交通运输中的关键实验室(北京交通大学),教育部) ; Frontiers Science Center for Smart High-speed Railway System, Beijing Jiaotong University(智能高速铁路系统前沿科学中心,北京交通大学)
专题命中 点云 :Gaussian Splatting(title,abstract);point cloud(title,abstract);分类 cs.CV
AI总结 PointGS通过3D高斯散射统一中间表示,解决3D点云与2D图像的语义一致性问题,实现无监督分割,优于现有方法,在ScanNet-V2和S3DIS上分别提升0.9%和2.8%的mIoU。
Comments Accepted by Computer Vision and Pattern Recognition (CVPR) 2026