SCANnotate++
收藏资源简介:
SCANnotate++数据集是由格拉茨技术大学视觉计算研究所和法国巴黎高等桥路学校等机构的研究人员创建的,该数据集为ScanNet++ v1数据集中的物体提供了高质量的CAD模型对齐和姿态标注。数据集包含了280个扫描场景中5290个物体的注释,是通过自动化的方法生成的,旨在降低人工标注成本并提高注释质量。该数据集可应用于三维场景理解任务,如点云补全和单视图CAD模型检索与对齐。
The SCANnotate++ dataset was developed by researchers from institutions including the Institute of Visual Computing of Graz University of Technology and École des Ponts ParisTech in Paris, France. This dataset offers high-quality CAD model alignment and pose annotations for objects within the ScanNet++ v1 dataset. Containing annotations for 5290 objects across 280 scanned scenes, the dataset was generated through automated methods, with the goal of lowering manual annotation costs and enhancing annotation quality. It can be applied to 3D scene understanding tasks such as point cloud completion and single-view CAD model retrieval and alignment.

- 1Leveraging Automatic CAD Annotations for Supervised Learning in 3D Scene Understanding格拉茨技术大学视觉计算研究所, 法国巴黎高等桥路学校, IP巴黎, CNRS · 2025年



