3D / Point Cloud
Objaverse
800K+ annotated 3D models for training 3D and multimodal models.
100K–1M·CC-BY·Point Cloud (LAS/PLY)
CloSe-D: A 3D Clothing Segmentation Dataset Project Page · Paper (arXiv:2401.12051) · Code (GitHub) CloSe-D is a large-scale 3D clothing segmentation dataset accompanying the 3DV 2024 paper “CloSe: A 3D Clothing Segmentation Dataset and Model”. It provides per-vertex clothing labels, SMPL body registrations, and canonical-pose annotations for thousands of clothed human scans across 18 garment categories. Dataset Subsets The dataset is released in three subsets with… See the full description on the dataset page:
Source: Hugging Face Hub (anticdimi/CloSe-D). Metadata imported from the dataset’s Hub tags.