Research map / 3D vision / Gaussian splatting
FillGS: Filling Observation Gaps in 4D Gaussian Splatting via Viewpoint-Time Selection and Generative Refinement
Takashi Otonari, Toshihiko Yamasaki
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 starting-viewpoint selection (ablation) | FillGS · Neural 3D Video interpolation | FID: 126.78 -> 83.81 | Improved |
| add sensitivity-based starting selection (ablation) | FillGS · Neural 3D Video interpolation | FID: 100.77 -> 83.81 | Improved |
| add error-based starting selection (ablation) | FillGS · Neural 3D Video interpolation | FID: 100.14 -> 83.81 | Improved |
| add candidate-viewpoint selection (ablation) | FillGS · Neural 3D Video interpolation | PSNR: 21.13 -> 21.73 | Improved |
| add observation-count-based selection (ablation) | FillGS · Neural 3D Video interpolation | PSNR: 20.55 -> 21.73 | Improved |
| add speed-based selection (ablation) | FillGS · Neural 3D Video interpolation | PSNR: 21.06 -> 21.73 | Improved |
| add global candidate view creation (ablation) | FillGS · Neural 3D Video extrapolation | PSNR: 15.85 -> 16.24 | Improved |
| use proposed score instead of negative (inverted) selection (ablation) | FillGS · Neural 3D Video interpolation | PSNR: 20.31 -> 21.73 | Improved |
| add weighting of generated images (ablation) | FillGS · Neural 3D Video interpolation | PSNR: 20.05 -> 21.73 | Improved |
| add consistency mask (ablation) | FillGS · Neural 3D Video interpolation | FID: 107.96 -> 83.81 | Improved |
| add co-visibility mask (ablation) | FillGS · Neural 3D Video interpolation | FID: 98.43 -> 83.81 | Improved |
| add co-visibility mask (ablation) | FillGS · Neural 3D Video interpolation | PSNR: 21.96 -> 21.73 | Hurt |
Closest papers by meaning
- Generative Sparse-View Gaussian Splatting (CVPR 2025 · similarity 0.92)
- 3DGS-Enhancer: Enhancing Unbounded 3D Gaussian Splatting with View-consistent 2D Diffusion Priors (NeurIPS 2024 · similarity 0.92)
- 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization (NeurIPS 2024 · similarity 0.91)
- Quantifying and Alleviating Co-Adaptation in Sparse-View 3D Gaussian Splatting (NeurIPS 2025 · similarity 0.91)
- SplatFlow: Multi-View Rectified Flow Model for 3D Gaussian Splatting Synthesis (CVPR 2025 · similarity 0.90)
- Taming Video Diffusion Prior with Scene-Grounding Guidance for 3D Gaussian Splatting from Sparse Inputs (CVPR 2025 · 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.
