OmniRig
收藏资源简介:
OmniRig是一个由清华大学深圳国际研究生院创建的大规模绑定数据集,包含79,499个具有详细骨架和蒙皮信息的网格模型。该数据集从多个来源收集和筛选而来,包括ModelResource、ObjaverseXL和互联网上公开可用的数据。数据集经过两阶段的过滤和后处理,以确保模型质量、多样性和完整性。OmniRig数据集覆盖了多种类别,包括复杂角色、简单角色、动物、海洋生物、鸟类、昆虫、植物等,适用于广泛的绑定应用。该数据集旨在促进三维绑定、姿态估计和动画合成等领域的研究。
OmniRig is a large-scale rigging dataset developed by Tsinghua University Shenzhen International Graduate School. It contains 79,499 mesh models with detailed skeleton and skinning information. The dataset is collected and filtered from multiple sources, including ModelResource, ObjaverseXL, and publicly available data on the Internet. It undergoes two-stage filtering and post-processing to guarantee model quality, diversity and integrity. The OmniRig dataset covers diverse categories such as complex characters, simple characters, animals, marine organisms, birds, insects and plants, and is applicable to a wide range of rigging-related applications. This dataset is intended to advance research in fields including 3D rigging, pose estimation and animation synthesis.

- 1ARMO: Autoregressive Rigging for Multi-Category Objects清华大学深圳国际研究生院, 中国 · 2025年




