LDM: Large Tensorial SDF Model for Textured Mesh Generation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866913543914258432 |
|---|---|
| author | Xie, Rengan Zheng, Wenting Huang, Kai Chen, Yizheng Wang, Qi Ye, Qi Chen, Wei Huo, Yuchi |
| author_facet | Xie, Rengan Zheng, Wenting Huang, Kai Chen, Yizheng Wang, Qi Ye, Qi Chen, Wei Huo, Yuchi |
| contents | Previous efforts have managed to generate production-ready 3D assets from text or images. However, these methods primarily employ NeRF or 3D Gaussian representations, which are not adept at producing smooth, high-quality geometries required by modern rendering pipelines. In this paper, we propose LDM, a novel feed-forward framework capable of generating high-fidelity, illumination-decoupled textured mesh from a single image or text prompts. We firstly utilize a multi-view diffusion model to generate sparse multi-view inputs from single images or text prompts, and then a transformer-based model is trained to predict a tensorial SDF field from these sparse multi-view image inputs. Finally, we employ a gradient-based mesh optimization layer to refine this model, enabling it to produce an SDF field from which high-quality textured meshes can be extracted. Extensive experiments demonstrate that our method can generate diverse, high-quality 3D mesh assets with corresponding decomposed RGB textures within seconds. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2405_14580 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | LDM: Large Tensorial SDF Model for Textured Mesh Generation Xie, Rengan Zheng, Wenting Huang, Kai Chen, Yizheng Wang, Qi Ye, Qi Chen, Wei Huo, Yuchi Graphics Previous efforts have managed to generate production-ready 3D assets from text or images. However, these methods primarily employ NeRF or 3D Gaussian representations, which are not adept at producing smooth, high-quality geometries required by modern rendering pipelines. In this paper, we propose LDM, a novel feed-forward framework capable of generating high-fidelity, illumination-decoupled textured mesh from a single image or text prompts. We firstly utilize a multi-view diffusion model to generate sparse multi-view inputs from single images or text prompts, and then a transformer-based model is trained to predict a tensorial SDF field from these sparse multi-view image inputs. Finally, we employ a gradient-based mesh optimization layer to refine this model, enabling it to produce an SDF field from which high-quality textured meshes can be extracted. Extensive experiments demonstrate that our method can generate diverse, high-quality 3D mesh assets with corresponding decomposed RGB textures within seconds. |
| title | LDM: Large Tensorial SDF Model for Textured Mesh Generation |
| topic | Graphics |
| url | https://arxiv.org/abs/2405.14580 |