Research map / 3D vision / 3D generation
UniGeo: Unifying Geometric Constraints for Camera-Controllable Image Editing via Video Priors
Hong Jiang, Wensong Song, Zongxin Yang, Ruijie Quan, Yi Yang
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 Frame-Decoupled Point Cloud Injection (ablation) | UniGeo · DL3DV (extensive) | FID: 121.52 -> 113.11 | Improved |
| add Frame-Decoupled Point Cloud Injection (ablation) | UniGeo · DL3DV (limited) | FID: 75.79 -> 69.05 | Improved |
| add Geometric Anchor Attention (ablation) | UniGeo · DL3DV (extensive) | FID: 115.83 -> 113.11 | Improved |
| add Geometric Anchor Attention (ablation) | UniGeo · DL3DV (limited) | FID: 70.44 -> 69.05 | Improved |
| add Trajectory-Endpoint Geometric Supervision (ablation) | UniGeo · DL3DV (extensive) | FID: 119.43 -> 113.11 | Improved |
| add Trajectory-Endpoint Geometric Supervision (ablation) | UniGeo · DL3DV (limited) | FID: 69.15 -> 69.05 | Improved |
| set GAA weight alpha to 0.1 instead of 1.0 (ablation) | UniGeo · DL3DV (extensive) | FID: 113.11 -> 115.17 | Hurt |
| set GAA weight alpha to 1.5 instead of 1.0 (ablation) | UniGeo · DL3DV (extensive) | FID: 113.11 -> 114.80 | Hurt |
| set TEGS gamma to 0.001 instead of 0.01 (ablation) | UniGeo · DL3DV (extensive) | FID: 113.11 -> 119.58 | Hurt |
| set TEGS gamma to 0.1 instead of 0.01 (ablation) | UniGeo · DL3DV (extensive) | FID: 113.11 -> 115.60 | Hurt |
| supervise only trajectory endpoints (no intermediate supervision) (ablation) | UniGeo · qualitative DL3DV | visual sharpness | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- ViewCrafter — ViewCrafter: Taming Video Diffusion Models for High-fidelity Novel View Synthesis (TPAMI 2025) · better on 2 of 2 shared results
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
- GeoNVS: Geometry Grounded Video Diffusion for Novel View Synthesis (ECCV 2026 · similarity 0.89)
- Video Camera Trajectory Editing with Generative Rendering from Estimated Geometry (AAAI 2026 · similarity 0.88)
- Geometrically Consistent Multi-View Scene Generation from Freehand Sketches (ECCV 2026 · similarity 0.88)
- Video Stitching from Multiple Moving Cameras (NeurIPS 2026 · similarity 0.88)
- SierpinskiCam: Camera-Controlled Video Retaking with Sierpinski Triangle Pattern Cues (NeurIPS 2026 · similarity 0.88)
- Less Decoder is More Encoder: Geometric Representation Learning from Novel View Synthesis (NeurIPS 2026 · 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.
