VAST-AI/NeuFrameQ-Dataset

NeuFrameQ: Neural Frame Fields for Scalable and Generalizable Anisotropic Quadrangulation (ICCV 2025)

 
Ying-Tian Liu¹, Jiajun Li³, Yu-Tao Liu², Xin Yu⁴, Yuan-Chen Guo², Yan-Pei Cao², Ding Liang², Ariel Shamir⁵, Song-Hai Zhang^1
1Tsinghua University · ²VAST · ³University of Chicago · ⁴The University of Hong Kong · ⁵Reichman University

---

Introduction

NeuFrameQ-Dataset is a large-scale collection of quadrilateral meshes released together with our paper NeuFrameQ: Neural Frame Fields for Scalable and Generalizable Anisotropic Quadrangulation (ICCV 2025).

The dataset consists of 279,680 quad meshes with diverse geometry, intended to support research in quadrangulation, remeshing, geometry processing, and so on. Importantly, the models in NeuFrameQ-Dataset are a subset of Objaverse. The UUIDs are consistent with Objaverse, ensuring easy cross-referencing between datasets.

• Models: There are 279,680 meshes stored in .ply format, grouped by the first two characters of their UUIDs.
• Metadata: Each group of models has a corresponding metadata file (.json.gz) located in the metadata/ directory. The metadata includes license information, UUIDs, and other attributes.

---

License

The dataset as a whole is released under the ODC-By v1.0 license.

Individual meshes are licensed under Creative Commons terms as provided in the metadata. Please consult the metadata files to verify the license of each object.

License Distribution:

• CC BY: 266,419
• CC BY-NC-SA: 2,088
• CC BY-SA: 2,445
• CC BY-NC: 7,858
• CC0: 870

---

BibTex

@inproceedings{liu2025neuframeq,
  title     = {NeuFrameQ: Neural Frame Fields for Scalable and Generalizable Anisotropic Quadrangulation},
  author    = {Liu, Ying-Tian and Li, Jiajun and Liu, Yu-Tao and Yu, Xin and Guo, Yuan-Chen and Cao, Yan-Pei and Liang, Ding and Shamir, Ariel and Zhang, Song-Hai},
  booktitle = {Proceedings of the IEEE/CVF International Conference on Computer Vision (ICCV)},
  year      = {2025}
}