Research map / 3D vision / 3D generation
Zero-1-to-3: Zero-shot One Image to 3D Object
Ruoshi Liu, Rundi Wu, Basile Van Hoorick, Pavel Tokmakov, Sergey Zakharov, Carl Vondrick
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| fine-tune Stable Diffusion with relative camera pose conditioning on Objaverse renders | Stable Diffusion · GSO novel view synthesis | PSNR: 8.933 -> 18.378 | Improved |
| apply view-conditioned model zero-shot to out-of-distribution multi-object scenes | Zero-1-to-3 · RTMV novel view synthesis | PSNR: 7.953 -> 10.405 | Improved |
| use view-conditioned diffusion as SJC prior for 3D reconstruction instead of Point-E / MCC | Zero-1-to-3 · GSO 3D reconstruction | volumetric IoU: 0.2944 -> 0.5052 | Improved |
| apply Zero-1-to-3 reconstruction to cluttered scenes | Zero-1-to-3 · RTMV 3D reconstruction | volumetric IoU: 0.1550 -> 0.2196 | Improved |
| reduce batch size from 1536 to 192 at 512x512 resolution (ablation) | Zero-1-to-3 fine-tuning · training convergence | convergence and batch variance | Hurt |
| reduce image size to 256x256 to allow batch size 1536 (ablation) | Zero-1-to-3 fine-tuning · training stability | convergence | Improved |
| add depth smoothness loss and near-view consistency loss to SJC | 3D reconstruction · 3D reconstruction (qualitative) | geometry and texture consistency | Improved |
Later papers that compared with this method
19 later accepted papers report Zero-1-to-3 as a baseline in their results tables; 18 of them report a better number on most of the results they share with it.
- NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis (NeurIPS 2026 · better on 10 of 10 shared results)
- Free-Form Scene Editor: Enabling Multi-Round Object Manipulation Like in a 3D Engine (AAAI 2026 · better on 6 of 6 shared results)
- MOVIS: Enhancing Multi-Object Novel View Synthesis for Indoor Scenes (CVPR 2025 · better on 10 of 10 shared results)
- DreamComposer++: Empowering Diffusion Models with Multi-View Conditions for 3D Content Generation (TPAMI 2025 · better on 7 of 7 shared results)
- Gamba: Marry Gaussian Splatting With Mamba for Single-View 3D Reconstruction (TPAMI 2025 · better on 3 of 3 shared results)
- V3D: Video Diffusion Models are Effective 3D Generators (TPAMI 2025 · better on 3 of 3 shared results)
- DreamComposer: Controllable 3D Object Generation via Multi-View Conditions (CVPR 2024 · better on 12 of 12 shared results)
- ViVid-1-to-3: Novel View Synthesis with Video Diffusion Models (CVPR 2024 · better on 10 of 10 shared results)
11 more not listed.
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- DiffPortrait3D: Controllable Diffusion for Zero-Shot Portrait View Synthesis (CVPR 2024 · similarity 0.91)
- ZeroNVS: Zero-Shot 360-Degree View Synthesis from a Single Image (CVPR 2024 · similarity 0.90)
- The More You See in 2D the More You Perceive in 3D (CVPR 2024 · similarity 0.90)
- ViVid-1-to-3: Novel View Synthesis with Video Diffusion Models (CVPR 2024 · similarity 0.90)
- Novel View Synthesis with Pixel-Space Diffusion Models (CVPR 2025 · similarity 0.90)
- Efficient-3Dim: Learning a Generalizable Single-image Novel-view Synthesizer in One Day (ICLR 2024 · similarity 0.90)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
