Research map / 3D vision / Gaussian splatting
FreeScale: Scaling 3D Scenes via Certainty-Aware Free-View Generation
Chenhan Jiang, Yu Chen, Qingwen Zhang, Jifei Song, Songcen Xu, Dit-Yan Yeung, Jiankang Deng
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 |
|---|---|---|---|
| train with FreeScale free-views | LVSM · DL3DV large camera motion | PSNR: 18.75 -> 21.45 | Improved |
| train with FreeScale free-views | LVSM · DL3DV small camera motion | PSNR: 22.20 -> 24.20 | Improved |
| train with FreeScale free-views | LVSM · Tanks & Temples zero-shot, small motion | PSNR: 13.07 -> 13.80 | Improved |
| add FreeScale free-views to per-scene 3DGS | 3DGS · DL3DV per-scene | PSNR: 19.18 -> 19.57 | Improved |
| add DIFIX3D enhancement to 3DGS | 3DGS · Nerfbusters per-scene | PSNR: 18.14 -> 17.69 | Hurt |
| add randomly sampled free-views (ablation) | LVSM · DL3DV large motion, 8K iterations | PSNR: 17.75 -> 18.68 | Improved |
| add view-graph guided training and selection (ablation) | LVSM + free-views · DL3DV large motion, 8K iterations | PSNR: 18.68 -> 19.11 | Improved |
| add view-graph selection to random free-views (ablation) | 3DGS per-scene · DL3DV per-scene | PSNR: 19.20 -> 19.57 | Improved |
| add view graph to certainty-based selection (ablation) | FreeScale · DL3DV large motion, 8K iterations | PSNR: 17.63 -> 19.11 | Improved |
| add certainty grid for view graph construction (ablation) | FreeScale · DL3DV large motion, 8K iterations | PSNR: 17.86 -> 19.11 | Improved |
| use view-graph reference instead of distance-based reference for rectification (ablation) | FreeScale per-scene · DL3DV per-scene | PSNR: 17.88 -> 19.57 | Improved |
| add diffusion-based image rectification (ablation) | LVSM + free-views · DL3DV large motion | PSNR: 20.89 -> 21.45 | Improved |
Closest papers by meaning
- FreeSplat: Generalizable 3D Gaussian Splatting Towards Free View Synthesis of Indoor Scenes (NeurIPS 2024 · similarity 0.90)
- Free3D: Consistent Novel View Synthesis without 3D Representation (CVPR 2024 · similarity 0.90)
- MegaSynth: Scaling Up 3D Scene Reconstruction with Synthesized Data (CVPR 2025 · similarity 0.89)
- NAMVIS: Next-Scale Autoregressive Multi-View Image Synthesis (NeurIPS 2026 · similarity 0.89)
- MVSplat360: Feed-Forward 360 Scene Synthesis from Sparse Views (NeurIPS 2024 · similarity 0.89)
- Novel View Synthesis from A Few Glimpses via Test-Time Natural Video Completion (NeurIPS 2025 · 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.
