Hunyuan3D-Omni: A Unified Framework for Controllable Generation of 3D Assets
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , , , , , , , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866915513889718272 |
|---|---|
| author | Hunyuan3D, Team : Zhang, Bowen Guo, Chunchao Liu, Haolin Yan, Hongyu Shi, Huiwen Huang, Jingwei Yu, Junlin Li, Kunhong Linus Wang, Penghao Lin, Qingxiang Liu, Sicong Yang, Xianghui Tang, Yixuan Zhao, Yunfei Lai, Zeqiang Liang, Zhihao Zhao, Zibo |
| author_facet | Hunyuan3D, Team : Zhang, Bowen Guo, Chunchao Liu, Haolin Yan, Hongyu Shi, Huiwen Huang, Jingwei Yu, Junlin Li, Kunhong Linus Wang, Penghao Lin, Qingxiang Liu, Sicong Yang, Xianghui Tang, Yixuan Zhao, Yunfei Lai, Zeqiang Liang, Zhihao Zhao, Zibo |
| contents | Recent advances in 3D-native generative models have accelerated asset creation for games, film, and design. However, most methods still rely primarily on image or text conditioning and lack fine-grained, cross-modal controls, which limits controllability and practical adoption. To address this gap, we present Hunyuan3D-Omni, a unified framework for fine-grained, controllable 3D asset generation built on Hunyuan3D 2.1. In addition to images, Hunyuan3D-Omni accepts point clouds, voxels, bounding boxes, and skeletal pose priors as conditioning signals, enabling precise control over geometry, topology, and pose. Instead of separate heads for each modality, our model unifies all signals in a single cross-modal architecture. We train with a progressive, difficulty-aware sampling strategy that selects one control modality per example and biases sampling toward harder signals (e.g., skeletal pose) while downweighting easier ones (e.g., point clouds), encouraging robust multi-modal fusion and graceful handling of missing inputs. Experiments show that these additional controls improve generation accuracy, enable geometry-aware transformations, and increase robustness for production workflows. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_21245 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Hunyuan3D-Omni: A Unified Framework for Controllable Generation of 3D Assets Hunyuan3D, Team : Zhang, Bowen Guo, Chunchao Liu, Haolin Yan, Hongyu Shi, Huiwen Huang, Jingwei Yu, Junlin Li, Kunhong Linus Wang, Penghao Lin, Qingxiang Liu, Sicong Yang, Xianghui Tang, Yixuan Zhao, Yunfei Lai, Zeqiang Liang, Zhihao Zhao, Zibo Computer Vision and Pattern Recognition Artificial Intelligence Recent advances in 3D-native generative models have accelerated asset creation for games, film, and design. However, most methods still rely primarily on image or text conditioning and lack fine-grained, cross-modal controls, which limits controllability and practical adoption. To address this gap, we present Hunyuan3D-Omni, a unified framework for fine-grained, controllable 3D asset generation built on Hunyuan3D 2.1. In addition to images, Hunyuan3D-Omni accepts point clouds, voxels, bounding boxes, and skeletal pose priors as conditioning signals, enabling precise control over geometry, topology, and pose. Instead of separate heads for each modality, our model unifies all signals in a single cross-modal architecture. We train with a progressive, difficulty-aware sampling strategy that selects one control modality per example and biases sampling toward harder signals (e.g., skeletal pose) while downweighting easier ones (e.g., point clouds), encouraging robust multi-modal fusion and graceful handling of missing inputs. Experiments show that these additional controls improve generation accuracy, enable geometry-aware transformations, and increase robustness for production workflows. |
| title | Hunyuan3D-Omni: A Unified Framework for Controllable Generation of 3D Assets |
| topic | Computer Vision and Pattern Recognition Artificial Intelligence |
| url | https://arxiv.org/abs/2509.21245 |