Research map / 3D vision / Gaussian splatting
Off The Grid: Detection of Primitives for Feed-Forward 3D Gaussian Splatting
Arthur Moreau, Richard Shaw, Michal Nazarczuk, Jisu Shin, Thomas Tanay, Zhensong Zhang, Songcen Xu, Eduardo Pérez-Pellitero
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 |
|---|---|---|---|
| replace voxel-aligned decoder with off-the-grid detected primitives | AnySplat (VGGT backbone) · NVS pose-free average | PSNR: 17.71 -> 21.21 | Improved |
| replace pixel-aligned decoder with off-the-grid detection (ablation) | Off The Grid · T&T + MipNeRF360 | LPIPS: 0.4341 -> 0.3796 | Improved |
| add predicted 3D offset to pixel-aligned Gaussians (SplatterImage) (ablation) | pixel-aligned baseline · T&T + MipNeRF360 | LPIPS: 0.4341 -> 0.4744 | Hurt |
| add self-rendering loss (ablation) | Off The Grid · T&T + MipNeRF360 | LPIPS: 0.4142 -> 0.3796 | Improved |
| add adaptive density (ablation) | Off The Grid · DL3DV benchmark | PSNR: 18.72 -> 19.09 | Improved |
| add confidence-based pruning (ablation) | Off The Grid · DL3DV benchmark | PSNR: 17.80 -> 19.09 | Improved |
| fine-tune VGGT with photometric loss through off-the-grid decoder | VGGT · Charge + SCRREAM | depth AbsRel: 0.14905 -> 0.1433; camera AUC@30 0.8957 -> 0.9278 | Mixed |
| fine-tune VGGT with AnySplat voxel decoder | VGGT · Charge + SCRREAM | camera pose AUC@30: 0.8957 -> 0.8333 | Hurt |
| use off-the-grid decoder instead of pixel-aligned DA3 GS | DA3 GS · NVS pose-free average | LPIPS: 0.3834 -> 0.3532 | Improved |
| use off-the-grid decoder with confidence pruning | AnySplat · primitive count | #G/pix: 0.8141 -> 0.1431 | 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.
- VGGT — VGGT: Visual Geometry Grounded Transformer (CVPR 2025) · better on 3 of 3 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
- Learning Efficient Fuse-and-Refine for Feed-Forward 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.92)
- CodecSplat: Ultra-Compact Latent Coding for Feed-Forward 3D Gaussian Splatting (NeurIPS 2026 · similarity 0.92)
- CompactSplat: Spatially Adaptive Gaussian Distribution for Feedforward Scene Reconstruction (NeurIPS 2026 · similarity 0.91)
- Speedy-Splat: Fast 3D Gaussian Splatting with Sparse Pixels and Sparse Primitives (CVPR 2025 · similarity 0.91)
- SfM-Free 3D Gaussian Splatting via Hierarchical Training (CVPR 2025 · similarity 0.91)
- KeyGS: A Keyframe-Centric Gaussian Splatting Method for Monocular Image Sequences (AAAI 2025 · similarity 0.91)
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.
