HPR3D: Hierarchical Proxy Representation for High-Fidelity 3D Reconstruction and Controllable Editing
Tielong Wang, Yuxuan Xiong, Jinfan Liu, Zhifan Zhang, Ye Chen, Yue Shi, Bingbing Ni · 2025
Read HPR3D: Hierarchical Proxy Representation for High-Fidelity 3D Reconstruction and Controllable Editing by Tielong Wang, Yuxuan Xiong, Jinfan Liu, Zhifan Zhang, Ye Chen, Yue Shi, Bingbing Ni (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-11971.txt.
Description
Wang and colleagues wrote in twenty twenty-five about a compact way to represent three-dimensional objects for reconstruction and editing. Meshes are precise but hard to edit, and neural fields render well but resist animation. Their representation stores shape and texture in a tree of proxy nodes. Nodes sit on the surface and inside the object, each covering a local patch. A small neural network encodes each patch's details implicitly. Querying any point blends nearby and parent node information.
The sparse node set keeps the representation compact overall. Nodes align with local parts, so dragging edits feel natural. Quality and size can be scaled by adjusting the tree. Experiments showed high-fidelity rendering with strong editability. The authors aim for one representation serving many tasks. Reconstruction, generation, and animation could share it, they argue.
Catalog record
| CREATOR | Tielong Wang, Yuxuan Xiong, Jinfan Liu, Zhifan Zhang, Ye Chen, Yue Shi, Bingbing Ni |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 35 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2507-11971