Research map / 3D vision / Gaussian splatting
FreeSplatter: Pose-free Gaussian Splatting for Sparse-view 3D Reconstruction
Jiale Xu, Shenghua Gao, Ying Shan
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 pixel-alignment loss (ablation) | FreeSplatter-O · GSO | PSNR: 26.684 -> 30.443 | Improved |
| add pixel-alignment loss (ablation) | FreeSplatter-S · ScanNet++ | PSNR: 21.330 -> 25.807 | Improved |
| add pixel-alignment loss (ablation) | FreeSplatter-S · ScanNet++ | LPIPS: 0.201 -> 0.140 | Improved |
| increase layers from 16 to 24 at patch size 8 (ablation) | FreeSplatter-O · GSO | PSNR: 28.945 -> 30.443 | Improved |
| increase layers from 16 to 24 at patch size 16 (ablation) | FreeSplatter-O · GSO | PSNR: 25.417 -> 28.622 | Improved |
| reduce patch size from 16 to 8 at 24 layers (ablation) | FreeSplatter-O · GSO | PSNR: 28.622 -> 30.443 | Improved |
| reduce patch size from 16 to 8 at 16 layers (ablation) | FreeSplatter-O · GSO | PSNR: 25.417 -> 28.945 | Improved |
| pose-free FreeSplatter-O versus pose-dependent LGM using GT poses | FreeSplatter-O · GSO | PSNR: 24.463 -> 30.443 | Improved |
| pose-free FreeSplatter-O versus pose-dependent InstantMesh using GT poses | FreeSplatter-O · OmniObject3D | PSNR: 24.077 -> 31.929 | Improved |
| FreeSplatter-O versus PF-LRM on PF-LRM's GSO eval data (GT novel-view poses) | FreeSplatter-O · GSO (PF-LRM eval data) | PSNR: 25.08 -> 23.54 | Hurt |
| FreeSplatter-O versus PF-LRM on PF-LRM's OmniObject3D eval data | FreeSplatter-O · OmniObject3D (PF-LRM eval data) | PSNR: 21.77 -> 22.83 | Improved |
| end-to-end FreeSplatter-S versus Splatt3R with frozen MASt3R | FreeSplatter-S · ScanNet++ | PSNR: 21.013 -> 25.807 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.
- PF-LRM — PF-LRM: Pose-Free Large Reconstruction Model for Joint Pose and Shape Prediction (ICLR 2024) · better on 2 of 4 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
- No Pose, No Problem: Surprisingly Simple 3D Gaussian Splats from Sparse Unposed Images (ICLR 2025 · similarity 0.95)
- SelfSplat: Pose-Free and 3D Prior-Free Generalizable 3D Gaussian Splatting (CVPR 2025 · similarity 0.92)
- PF3plat: Pose-Free Feed-Forward 3D Gaussian Splatting for Novel View Synthesis (ICML 2025 · similarity 0.92)
- FreeSplat: Generalizable 3D Gaussian Splatting Towards Free View Synthesis of Indoor Scenes (NeurIPS 2024 · similarity 0.91)
- OnlineSplatter: Pose-Free Online 3D Reconstruction for Free-Moving Objects (NeurIPS 2025 · similarity 0.91)
- pixelSplat: 3D Gaussian Splats from Image Pairs for Scalable Generalizable 3D Reconstruction (CVPR 2024 · similarity 0.90)
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.
