Research map / 3D vision / Gaussian splatting
MSCD-GS: Motion-Separated Cooperative Deblurring Dynamic Reconstruction via Gaussian Splatting
Yongjian Liao, Xu Zou, Wenjun Chen, Huixuan Li, Xiaoen Xie, Chunxi Li, Shixiang Huang, Gang Zhang, Jiahuan Zhou, Sheng Zhong, Luxin Yan
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 NAFNet-supervised SoM with MSCD-GS cooperative supervision | SoM + NAFNet · Stereo Blur, novel view synthesis | PSNR: 29.49 -> 27.76 (MSCD-GS vs SoM + NAFNet) | Improved |
| add MSCD-GS cooperative deblurring on top of deblurring model | NAFNet · Stereo Blur, deblurring | PSNR: 32.65 -> 33.21 | Improved |
| add pre-trained deblurring network supervision (ablation) | MSCD-GS · Stereo Blur Man scene | PSNR | Improved |
| add dynamic Gaussian deblurring module (ablation) | MSCD-GS · Stereo Blur Man scene | PSNR | Improved |
| combine DN, static Gaussian deblurring and dynamic Gaussian deblurring (all modules) (ablation) | MSCD-GS · Stereo Blur Man scene | deblurring PSNR: 31.72 (full) | Improved |
| use 3 virtual views instead of 2 (ablation) | MSCD-GS · Stereo Blur Man scene | deblurring PSNR: 31.32 -> 31.72 | Improved |
| use 4 virtual views instead of 3 (ablation) | MSCD-GS · Stereo Blur Man scene | deblurring PSNR: 31.72 -> 31.81 | No clear effect |
| increase virtual views to 6 and 10 (ablation) | MSCD-GS · Stereo Blur Man scene | deblurring PSNR: 31.72 -> 31.12 (6), 30.72 (10) | Hurt |
| increase virtual views from 3 to 10 (ablation) | MSCD-GS · Stereo Blur Man scene | training time (h): 0.72 -> 2.51 | Hurt |
| MSCD-GS vs prior dynamic deblurring methods on real-world blurry data | BARD-GS · Real-world blurry dataset, novel view synthesis | PSNR: 25.13 -> 28.13 | Improved |
| add NAFNet deblurring to E-D3DGS | E-D3DGS · Stereo Blur, deblurring | PSNR: 27.68 -> 28.92 | Improved |
| add NAFNet deblurring to SoM | SoM · Stereo Blur, novel view synthesis | PSNR: 26.87 -> 27.76 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- Deblur4DGS — Deblur4DGS: 4D Gaussian Splatting from Blurry Monocular Video (AAAI 2026) · better on 8 of 8 shared results
- BARD-GS — BARD-GS: Blur-Aware Reconstruction of Dynamic Scenes via Gaussian Splatting (CVPR 2025) · better on 6 of 6 shared results
- DyBluRF — DyBluRF: Dynamic Neural Radiance Fields from Blurry Monocular Video (CVPR 2024) · 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
- BARD-GS: Blur-Aware Reconstruction of Dynamic Scenes via Gaussian Splatting (CVPR 2025 · similarity 0.93)
- Motion Decoupled 3D Gaussian Splatting for Dynamic Object Representation (AAAI 2025 · similarity 0.91)
- GS-Blur: A 3D Scene-Based Dataset for Realistic Image Deblurring (NeurIPS 2024 · similarity 0.91)
- Dynamic Gaussian Splatting from Defocused and Motion-blurred Monocular Videos (NeurIPS 2025 · similarity 0.90)
- 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization (NeurIPS 2024 · similarity 0.90)
- DiET-GS: Diffusion Prior and Event Stream-Assisted Motion Deblurring 3D Gaussian Splatting (CVPR 2025 · similarity 0.89)
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.
