Research map / 3D vision / Gaussian splatting
PRISM3D: Probabilistic Refinement and Robust Initialization for Physically Consistent Scene Modeling under Extreme Motion Blur
Gopi Raju Matta, Reddypalli Trisha, Divya Madhuri Vemunuri, Kaushik Mitra
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 |
|---|---|---|---|
| use VGGSfM instead of HLOC for initialization (ablation) | PRISM3D · Synthetic ablation (8 scenes) | PSNR: 28.04 -> 29.49; 1.45 dB | Improved |
| use VGGSfM instead of VGGT-X for initialization (ablation) | PRISM3D · Synthetic ablation (8 scenes) | PSNR: 27.82 -> 29.49; 1.67 dB | Improved |
| add MCMC-based probabilistic refinement (ablation) | PRISM3D · Synthetic ablation (8 scenes) | PSNR: 29.02 -> 29.49; 0.47 dB | Improved |
| add MCMC refinement (ablation) | PRISM3D · Synthetic ablation (8 scenes) | LPIPS: 0.133 -> 0.153 | Hurt |
| use Bezier trajectory instead of linear (ablation) | PRISM3D · Synthetic ablation (8 scenes) | PSNR: 27.94 -> 29.49; 1.55 dB | Improved |
| use Bezier trajectory instead of spline (ablation) | PRISM3D · Synthetic ablation (8 scenes) | PSNR: 29.25 -> 29.49; 0.24 dB | Improved |
| use Bezier trajectory instead of linear and spline (ablation) | PRISM3D-E · Synthetic ablation (8 scenes) | PSNR: 29.77 -> 31.95; 2.18 dB and 0.59 dB | Improved |
| add event-derived priors to initialization (PRISM3D-E) (ablation) | PRISM3D · Synthetic ablation (8 scenes) | PSNR: 29.49 -> 31.95; 2.46 dB | Improved |
| add MCMC refinement (ablation) | PRISM3D-E · Synthetic ablation (8 scenes) | PSNR: 31.06 -> 31.95 | Improved |
| use VGGSfM instead of COLMAP for initialization on EDI images (ablation) | PRISM3D-E · Synthetic ablation (8 scenes) | PSNR: 31.28 -> 31.95 | Improved |
| add explicit event reconstruction loss during optimization (ablation) | PRISM3D-E · Synthetic ablation (8 scenes) | PSNR: 31.95 -> 31.86 | No clear effect |
| increase virtual cameras n from 5 to 15 (ablation) | PRISM3D · Synthetic trajectory analysis | PSNR | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- BAD-Gaussians* — BAD-Gaussians: Bundle Adjusted Deblur Gaussian Splatting (ECCV 2024) · better on 4 of 5 shared results
- ExBluRF* — ExBluRF: Efficient Radiance Fields for Extreme Motion Blurred Images (ICCV 2023) · better on 1 of 1 shared result
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
- Splat-based 3D Scene Reconstruction with Extreme Motion-blur (ICCV 2025 · similarity 0.89)
- DGS-LRM: Real-Time Deformable 3D Gaussian Reconstruction From Monocular Videos (NeurIPS 2025 · similarity 0.87)
- Motion4D: Learning 3D-Consistent Motion and Semantics for 4D Scene Understanding (NeurIPS 2025 · similarity 0.87)
- Coherent 3D Scene Diffusion From a Single RGB Image (NeurIPS 2024 · similarity 0.87)
- HiCoM: Hierarchical Coherent Motion for Dynamic Streamable Scenes with 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.87)
- Metropolis-Hastings Sampling for 3D Gaussian Reconstruction (NeurIPS 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.
