Research map / 3D vision / 3D generation
GeoNVS: Geometry Grounded Video Diffusion for Novel View Synthesis
Minjun Kang, Inkyu Shin, Taeyeop Lee, Myungchul Kim, In Kweon, KUK-JIN YOON
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 |
|---|---|---|---|
| add GS-Adapter feature-space geometry guidance to SEVA | SEVA · Small-viewpoint set (average) | PSNR: 17.15 -> 19.09 | Improved |
| add GS-Adapter to SEVA on large-viewpoint set | SEVA · Large-viewpoint set (average) | PSNR: 12.64 -> 14.20 | Improved |
| add GS-Adapter to SEVA for long trajectories | SEVA · Long trajectory DL3DV (P=9) | PSNR: 16.73 -> 19.32 | Improved |
| add GS-Adapter to SEVA, camera pose and geometry | SEVA · RE10K (P=3) | Translation error: 1.62 -> 0.76 | Improved |
| add GS-Adapter to SEVA, Chamfer distance | SEVA · RE10K (P=3) | Chamfer distance: 1.35 -> 0.19 | Improved |
| replace feature-space guidance with input-level injection of 3D Gaussian renders | SEVA · DL3DV (6 reference views) | Translation error: 19.68 -> 81.14 | Hurt |
| use Difix3D input-level geometry integration instead of GS-Adapter | SEVA · DL3DV (6 reference views) | Chamfer distance: 0.57 -> 3.76 | Hurt |
| add GS-Adapter to improve unseen-region quality | SEVA · DL3DV (6 reference views) | PSNR_U: +2.08 | Improved |
| use adaptive fusion instead of naive fusion with Pi3 prior (ablation) | GeoNVS (SEVA) · DTU (3 reference views) | PSNR: 14.83 -> 19.02 | Improved |
| use adaptive fusion instead of naive fusion with VGGT prior (ablation) | GeoNVS (SEVA) · DTU (3 reference views) | PSNR: 18.89 -> 19.18 | Improved |
| add adaptive fusion (vs naive fusion) (ablation) | GS-Adapter (VGGT prior) · Mip360 6-view | PSNR: 16.48 -> 16.57 | Improved |
| add feature consistency loss L_feat (ablation) | GS-Adapter (VGGT prior) · Mip360 6-view | PSNR: 16.47 -> 16.48 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 5 on most shared results.
- DepthSplat — DepthSplat: Connecting Gaussian Splatting and Depth (CVPR 2025) · better on 2 of 2 shared results
- GenFusion — GenFusion: Closing the Loop between Reconstruction and Generation via Videos (CVPR 2025) · better on 2 of 2 shared results
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 1 of 2 shared results
- ViewCrafter — ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View Synthesis (TPAMI 2025) · better on 1 of 1 shared result
- MVSplat360 — MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views (NeurIPS 2024) · better on 1 of 1 shared result
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
- SpatialCrafter: Unleashing the Imagination of Video Diffusion Models for Scene Reconstruction from Limited Observations (ICCV 2025 · similarity 0.90)
- Repurposing Geometric Foundation Models for Multi-view Diffusion (ECCV 2026 · similarity 0.90)
- Novel View Synthesis with Pixel-Space Diffusion Models (CVPR 2025 · similarity 0.89)
- ViVid-1-to-3: Novel View Synthesis with Video Diffusion Models (CVPR 2024 · similarity 0.89)
- Enhancing Novel View Synthesis via Geometry Grounded Set Diffusion (NeurIPS 2026 · similarity 0.89)
- Geometry-guided Online 3D Video Synthesis with Multi-View Temporal Consistency (CVPR 2025 · similarity 0.89)
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.
