Research map / 3D vision / 3D generation
Zero-Shot Novel Depth Synthesis Using Foundation Models Scene Representations
Denis Akola, David Fouhey
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 |
|---|---|---|---|
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-VGGT · In-domain (1 source -> 1 target) | AbsRel: 0.098 -> 0.076 | Improved |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-WM · In-domain (1 source -> 1 target) | AbsRel: 0.108 -> 0.075 | Improved |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-VGGT · In-domain (2 sources -> 4 targets) | AbsRel: 0.248 -> 0.112 | Improved |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-WM · In-domain (2 sources -> 4 targets) | AbsRel: 0.215 -> 0.118 | Improved |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-VGGT · MegaDepth (2 sources -> 4 targets) | AbsRel: 0.102 -> 0.039 | Improved |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-VGGT · DTU (1 source -> 1 target) | AbsRel: 0.022 -> 0.012 | Improved |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-VGGT · NRGBD (1 source -> 1 target) | AbsRel: 0.192 -> 0.260 | Hurt |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-WM · NRGBD (1 source -> 1 target) | AbsRel: 0.127 -> 0.133 | No clear effect |
| diffuse in 3DFM latent space instead of depth space (ablation) | Z3D-VGGT · 7-Scenes (2 sources -> 4 targets) | AbsRel: 0.372 -> 0.342 | Improved |
| use Z3D latent diffusion instead of LVSM plus 3DFM depth | Z3D-VGGT · DTU (1 source -> 1 target) | AbsRel: 0.468 -> 0.012 | Improved |
| use Z3D instead of LVSM plus 3DFM | Z3D-WM · NRGBD (2 sources -> 4 targets) | AbsRel: 0.121 -> 0.131 | Mixed |
| linear probe 3DFM features for layered depth | VGGT features · LDI prediction | AbsRel: 0.319 -> 0.197 | Improved |
Closest papers by meaning
- Zero-Shot Novel View and Depth Synthesis with Multi-View Geometric Diffusion (CVPR 2025 · similarity 0.88)
- Probing the 3D Awareness of Visual Foundation Models (CVPR 2024 · similarity 0.88)
- UniSDF: Unifying Neural Representations for High-Fidelity 3D Reconstruction of Complex Scenes with Reflections (NeurIPS 2024 · similarity 0.87)
- VGGT: Visual Geometry Grounded Transformer (CVPR 2025 · similarity 0.87)
- BetterDepth: Plug-and-Play Diffusion Refiner for Zero-Shot Monocular Depth Estimation (NeurIPS 2024 · similarity 0.86)
- Efficient-3Dim: Learning a Generalizable Single-image Novel-view Synthesizer in One Day (ICLR 2024 · similarity 0.86)
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.
