Research map / 3D vision / Gaussian splatting
FixAnything: 3D-Consistent Rendering Refinement via Video Generative Priors
Khiem Vuong, Deva Ramanan, Srinivasa G. Narasimhan
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 Flow-DPO with SfM pose-accuracy reward (ablation) | FixAnything · DL3DV (6 views) | AUC@5deg: 61.12 -> 68.32 | Improved |
| add Flow-DPO (ablation) | FixAnything · DL3DV (6 views) | PSNR: 17.51 -> 17.65 | Improved |
| add binary clean-frame mask conditioning (ablation) | FixAnything · DL3DV (6 views) | PSNR: 16.37 -> 17.65 | Improved |
| train on 500 instead of 20 paired videos (ablation) | FixAnything · DL3DV (6 views) | PSNR: 16.70 -> 17.65 | Improved |
| train on 500 instead of 100 paired videos (ablation) | FixAnything · DL3DV (6 views) | PSNR: 17.45 -> 17.65 | Improved |
| reduce denoising steps from 50 to 5 (ablation) | FixAnything · DL3DV (6 views) | PSNR: 17.65 -> 18.02 | No clear effect |
| reduce denoising steps from 50 to 5 (ablation) | FixAnything · DL3DV (6 views) | runtime (s): 309 -> 31 | Improved |
| clean mesh rendering input vs 3DGS rendering input (ablation) | FixAnything · DL3DV (3 views) | PSNR: 15.18 -> 15.74 | Improved |
| use NeRF rendering instead of 3DGS rendering input (ablation) | FixAnything · DL3DV (3 views) | PSNR: 15.18 -> 14.22 | Hurt |
| use FixAnything on 3DGS rendering vs Difix3D+ | FixAnything · DL3DV (6 views) | PSNR: 14.41 -> 17.65 | Improved |
| use FixAnything on 3DGS rendering vs Xu et al. | FixAnything · DL3DV (9 views) | PSNR: 19.19 -> 19.76 | Improved |
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.
- 3DGS-Enhancer — 3DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priors (NeurIPS 2024) · better on 5 of 5 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
- MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views (NeurIPS 2024 · similarity 0.89)
- SimVS: Simulating World Inconsistencies for Robust View Synthesis (CVPR 2025 · similarity 0.89)
- GenFusion: Closing the Loop between Reconstruction and Generation via Videos (CVPR 2025 · similarity 0.89)
- Novel View Synthesis from A Few Glimpses via Test-Time Natural Video Completion (NeurIPS 2025 · similarity 0.89)
- UMAMI: Unifying Masked Autoregressive Models and Deterministic Rendering for View Synthesis (NeurIPS 2025 · similarity 0.88)
- GenRec: Knowing Where to Reconstruct and Where to Generate (NeurIPS 2026 · similarity 0.88)
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.
