PartUV: Part-Based UV Unwrapping of 3D Meshes

Zhaoning Wang1, Xinyue Wei1, 2, Ruoxi Shi1, 2, Xiaoshuai Zhang1, Hao Su1, 2, Minghua Liu1
1Hillbot Inc. 2University of California, San Diego
SIGGRAPH Asia 2025
Paper arXiv Video Code
PartUV method overview

We propose PartUV, a novel part-based UV unwrapping method for 3D meshes. Unlike traditional approaches that rely only on local geometric priors and often produce over-fragmented charts, PartUV combines learned part priors with geometric cues to generate a compact set of part-aligned charts. We evaluate our method on four diverse datasets, and show that it produces significantly less fragmented UV mappings while maintaining low distortion comparable to baseline methods. Leveraging part-aware charts further enables applications such as generating one UV atlas per semantic part.

Video

Abstract

UV unwrapping flattens 3D surfaces to 2D with minimal distortion, often requiring the complex surface to be decomposed into multiple charts. Although extensively studied, existing UV unwrapping methods frequently struggle with AI-generated or reconstructed meshes, which are typically noisy, bumpy, and poorly conditioned. These methods often produce highly fragmented charts and suboptimal boundaries, introducing artifacts and hindering downstream tasks. We introduce PartUV, a part-based UV unwrapping pipeline that generates significantly fewer, part-aligned charts while maintaining low distortion. Built on top of a recent learning-based part decomposition method PartField, PartUV combines high-level semantic part decomposition with novel geometric heuristics in a top-down recursive framework. It ensures each chart’s distortion remains below a user-specified threshold while minimizing the total number of charts. The pipeline integrates and extends parameterization and packing algorithms, incorporates dedicated handling of non-manifold and degenerate meshes, and is extensively parallelized for efficiency. Evaluated across four diverse datasets—including man-made, CAD, AI-generated, and Common Shapes—PartUV outperforms existing tools and recent neural methods in chart count and seam length, achieves comparable distortion, exhibits high success rates on challenging meshes, and enables new applications like part-specific multi-tiles packing.


Results

Our part-based method not only achieves significantly fewer charts while maintaining low distortion, but also enables part-based organization, where charts from the same part are packed near each other for better clarity.

For a detailed view of the results, please click here.


Method Overview

PartUV pipeline

Given a mesh M, we first apply the learning-based method PartField to predict a part-aware feature field. Clustering this field produces a hierarchical part tree T for the input mesh. We then recursively traverse T starting from the root. For each visited node P, we apply two geometry-based strategies to segment the corresponding part mesh into a set of charts C. Each chart is flattened using the ABF algorithm, and its distortion is evaluated. If the distortion exceeds a user-specified threshold τ, we recurse into the left and right children of T; otherwise, we adopt the segmented charts and their UV mappings for that part mesh.

Applications and Additional Comparisons

Please swipe or use the arrows to browse five example applications.

BibTeX

@inproceedings{wang2025partuv,
  title     = {PartUV: Part-Based UV Unwrapping of 3D Meshes},
  author    = {Wang, Zhaoning and Wei, Xinyue and Shi, Ruoxi and Zhang, Xiaoshuai and Su, Hao and Liu, Minghua},
  booktitle = {ACM SIGGRAPH Asia Conference and Exhibition on Computer Graphics and Interactive Techniques},
  year      = {2025}
}