Research map / 3D vision / 3D generation
SyncFix: Multi-View Consistent Diffusion Refinement of 3D Reconstructions
Deming Li, Cheng Peng, Abhay Kumar Yadav, Rama Chellappa, Anand Bhattad
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 joint multi-view refinement (cross-view attention) (ablation) | SyncFix (no reference) · DL3DV test (ablation) | CVSC: 0.821 -> 0.862 | Improved |
| add joint multi-view refinement (ablation) | SyncFix (no reference) · DL3DV test (ablation) | PSNR: 16.26 -> 16.45 | Improved |
| add joint multi-view refinement (ablation) | SyncFix (no reference) · DL3DV test (ablation) | LPIPS: 0.347 -> 0.334 | Improved |
| add reference views (ablation) | SyncFix · DL3DV test (ablation) | PSNR: 16.45 -> 16.94 | Improved |
| add reference views (ablation) | SyncFix · DL3DV test (ablation) | CVSC: 0.862 -> 0.880 | Improved |
| add reference views (ablation) | SyncFix · DL3DV test (ablation) | DreamSim: 0.129 -> 0.099 | Improved |
| give Difix3D+ multiple views at inference (ablation) | Difix3D+ · DL3DV test (ablation) | PSNR: 16.17 -> 16.08 | Hurt |
| give Difix3D+ multiple views at inference (ablation) | Difix3D+ · DL3DV test (ablation) | CVSC: 0.850 -> 0.827 | Hurt |
| use SyncFix latent bridge matching instead of Difix3D+ (single-view) (ablation) | SyncFix single-view · DL3DV test (ablation) | PSNR: 16.17 -> 16.26 | Improved |
| replace Difix3D+ with SyncFix | Difix3D+ · DL3DV test | PSNR: 16.17 -> 16.94 | Improved |
| replace Difix3D+ with SyncFix | Difix3D+ · DL3DV test | LPIPS: 0.343 -> 0.305 | Improved |
| replace Difix3D+ with SyncFix | Difix3D+ · DL3DV test | runtime: 30.6 -> 5.63 | 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.
- FlowR — FlowR: Flowing from Sparse to Dense 3D Reconstructions (ICCV 2025) · better on 7 of 7 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
- Carve3D: Improving Multi-view Reconstruction Consistency for Diffusion Models with RL Finetuning (CVPR 2024 · similarity 0.88)
- DIFIX3D+: Improving 3D Reconstructions with Single-Step Diffusion Models (CVPR 2025 · similarity 0.88)
- MultiDiff: Consistent Novel View Synthesis from a Single Image (CVPR 2024 · similarity 0.88)
- Decompositional Neural Scene Reconstruction with Generative Diffusion Prior (CVPR 2025 · similarity 0.88)
- ViewFusion: Towards Multi-View Consistency via Interpolated Denoising (CVPR 2024 · similarity 0.88)
- SyncNoise: Geometrically Consistent Noise Prediction for Instruction-based 3D Editing (AAAI 2025 · similarity 0.87)
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.
