Research map / 3D vision / Gaussian splatting
CoMoGaussian: Continuous Motion-Aware Gaussian Splatting from Motion-Blurred Images
Jungho Lee, Donghyeong Kim, Dogyoon Lee, Suhwan Cho, Minhyeok Lee, Wonjoon Lee, Taeoh Kim, Dongyoon Wee, Sangyoun Lee
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 rigid body transformation to Mip-Splatting baseline (ablation) | Mip-Splatting · Deblur-NeRF real-world | PSNR: 21.87 -> 27.19 | Improved |
| add CMR transformation on top of rigid motion (ablation) | CoMoGaussian · Deblur-NeRF real-world | PSNR: 27.19 -> 27.74 | Improved |
| add orthogonality loss on CMR rotation (ablation) | CoMoGaussian · Deblur-NeRF real-world | LPIPS: 0.0878 -> 0.0822 | Improved |
| replace linear interpolation / cubic B-spline physical-space trajectories with latent-space neural ODE (ablation) | CoMoGaussian · Deblur-NeRF real-world | PSNR: 21.69 -> 27.85 | Improved |
| replace MLP or GRU with neural ODE for latent features (ablation) | CoMoGaussian · Deblur-NeRF real-world | LPIPS: 0.1010 (MLP), 0.0988 (GRU) -> 0.0822 | Improved |
| add CMR so fewer poses are needed (ablation) | CoMoGaussian · Deblur-NeRF real-world | PSNR: 27.57 (rigid only, N=13) -> 27.85 (rigid+CMR, N=9) | Improved |
| increase number of poses N for rigid-only trajectory (ablation) | Rigid only · Deblur-NeRF real-world | PSNR: 26.60 -> 27.57 (N=5 to 13) | Improved |
| increase poses N beyond 9 with CMR (ablation) | Rigid + CMR · Deblur-NeRF real-world | PSNR: 27.85 (N=9) vs 27.89 (N=10) vs 27.81 (N=13) | No clear effect |
| remove orthogonal regularization (ablation) | CoMoGaussian · Deblur-NeRF real-world | visual detail | Improved |
| apply deblurring method on sharp images | CoMoGaussian vs Deblurring 3DGS · NeRF-LLFF | PSNR: CoMoGaussian 27.56 vs Mip-Splatting 27.71; Deblurring 3DGS 17.42 | Mixed |
| CoMoGaussian vs other deblurring 3DGS methods on training time | CoMoGaussian · Deblur-NeRF real-world | Training time (hours): 0.20 (Deblurring 3DGS) -> 1.33 | Hurt |
Later papers that compared with this method
1 later accepted paper reports CoMoGaussian as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Unblur-SLAM: Dense Neural SLAM for Blurry Inputs (CVPR 2026 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 4 on most shared results.
- BAD-Gaussians — BAD-Gaussians: Bundle Adjusted Deblur Gaussian Splatting (ECCV 2024) · better on 3 of 3 shared results
- Mip-Splatting — Mip-Splatting: Alias-free 3D Gaussian Splatting (CVPR 2024) · better on 1 of 2 shared results
- BAGS — BAGS: Blur Agnostic Gaussian Splatting through Multi-Scale Kernel Modeling (ECCV 2024) · better on 2 of 2 shared results
- DP-NeRF — DP-NeRF: Deblurred Neural Radiance Field With Physical Scene Priors (CVPR 2023) · better on 2 of 2 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
- MotionGS: Exploring Explicit Motion Guidance for Deformable 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.91)
- HiCoM: Hierarchical Coherent Motion for Dynamic Streamable Scenes with 3D Gaussian Splatting (NeurIPS 2024 · similarity 0.91)
- Dynamic Gaussian Splatting from Defocused and Motion-blurred Monocular Videos (NeurIPS 2025 · similarity 0.91)
- Deformable 3D Gaussians for High-Fidelity Monocular Dynamic Scene Reconstruction (CVPR 2024 · similarity 0.91)
- DGS-LRM: Real-Time Deformable 3D Gaussian Reconstruction From Monocular Videos (NeurIPS 2025 · similarity 0.91)
- CoCoGaussian: Leveraging Circle of Confusion for Gaussian Splatting from Defocused Images (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.
