Harnessing Generative Image Models for Training-Free Primitive Shape Abstraction
Quick Answer
This study introduces a training-free method for 3D shape abstraction using generative image models, achieving the lowest Chamfer distance on HumanPrim and Toys4K datasets with 5-9 primitives per object.
Quick Take
This study introduces a training-free method for 3D shape abstraction using generative image models, achieving the lowest Chamfer distance on HumanPrim and Toys4K datasets with 5-9 primitives per object. The approach leverages a for semantic analysis and is category-agnostic, addressing limitations of previous learning-based models.
Key Points
- Generative image models enable direct shape abstraction without fine-tuning.
- Achieves lowest Chamfer distance among evaluated methods on HumanPrim and Toys4K.
- Utilizes 5-9 geometric primitives per object for effective representation.
- Category-agnostic and orientation-invariant, overcoming previous model limitations.
- Part segmentation accuracy is the current bottleneck, not primitive fitting.
Paper Resources
📖 Reader Mode
~2 min readAbstract:Representing 3D shapes as compact sets of geometric primitives is fundamental to robotics, simulation, and scene understanding. Generative image models trained at scale have recently emerged as generalist visual learners that can identify and segment object parts directly in the image domain, across arbitrary categories and without task-specific training. Adapting such models to downstream tasks typically requires fine-tuning; we ask whether their pretrained capability can instead be harnessed directly, without any training, and answer affirmatively with a training-free harness. Our pipeline renders multi-view images of a 3D object, uses a vision-language model to analyze its semantic parts, prompts a generative image model to paint a color-coded part segmentation mask, reprojects it onto the geometry, and fits a superquadric primitive to each part via parameter optimization. The approach contains no learned parameters: it is category-agnostic and orientation-invariant, properties that previous learning-based models struggled with. Its accuracy ceiling rises with future generative-model improvements, which we confirm with a ground-truth segmentation study showing that part segmentation, not primitive fitting, is the current accuracy bottleneck. On HumanPrim and Toys4K, our method achieves the lowest Chamfer distance among all evaluated methods, using 5--9 primitives per object on average.
| Comments: | 13 pages, 9 figures, 3 tables |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2607.05568 [cs.CV] |
| (or arXiv:2607.05568v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2607.05568 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Gregor Kobsik [view email]
[v1]
Mon, 6 Jul 2026 19:10:43 UTC (13,471 KB)
— Originally published at arxiv.org
Want this in your inbox every morning?
Daily brief at your local 8am — bilingual EN/中文, free.
More from arXiv cs.CV
See more →LLM-Guided ANN Index Optimization for Human-Object Interaction Retrieval
A phase-aware LLM agent optimizes human-object interaction retrieval, outperforming Optuna TPE by 33.3% and VDTuner by 34.2% on the HICO-DET benchmark. This method enhances throughput by 15.3x over UniIR and demonstrates strong transferability across vector database management systems.