Research map / 3D vision / Gaussian splatting
$L^{2}DGS$: Low-Light Dynamic Gaussian Splatting
Ashish Kumar, Rajagopalan N Ambasamduram
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 Occlusion-Disocclusion Network (OCD-Net) (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 13.36 -> 14.68 | Improved |
| add BAFE-Net to enhance Brightness Attenuation Features (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 6.14 -> 14.68 | Improved |
| add signal amplification regularizer (maximize Lw) (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 6.99 -> 14.68 | Improved |
| add exposure regularizer on reflectance (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 7.65 -> 14.68 | Improved |
| add SSIM term to photometric loss (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 13.20 -> 14.68 | Improved |
| add edge-aware depth smoothness loss (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 14.27 -> 14.68 | Improved |
| add SASR gradient regularizers on B1 and B2 (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 14.56 -> 14.68 | Improved |
| also maximize reflectance Rw (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 14.68 -> 14.57 | No clear effect |
| maximize Rw instead of Lw (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 14.68 -> 6.89 | Hurt |
| fix B1 = 1 or B2 = 1 instead of learned BAFs (ablation) | L2DGS · Synthetic average (Sriracha, Creeper, Mochi, Apple), overall | PSNR: 14.68 -> 12.46 (B1=1), 12.54 (B2=1) | Hurt |
| compare L2DGS against baselines on Handwavy synthetic scene | L2DGS · Handwavy (synthetic, overall) | PSNR: 11.72 vs 12.78 (LLNeRF) | Mixed |
| L2DGS vs best baseline on Mochi dynamic-aware reconstruction | L2DGS · Mochi (synthetic, overall) | PSNR: 14.95 -> 17.61 | 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.
- Luminance-GS — Luminance-GS: Adapting 3D Gaussian Splatting to Challenging Lighting Conditions with View-Adaptive Curve Adjustment (CVPR 2025) · better on 3 of 4 shared results
- LITA-GS — LITA-GS: Illumination-Agnostic Novel View Synthesis via Reference-Free 3D Gaussian Splatting and Physical Priors (CVPR 2025) · better on 3 of 3 shared results
- AlethNeRF — Aleth-NeRF: Illumination Adaptive NeRF with Concealing Field Assumption (AAAI 2024) · better on 2 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
- DGS-LRM: Real-Time Deformable 3D Gaussian Reconstruction From Monocular Videos (NeurIPS 2025 · similarity 0.90)
- Efficient Gaussian Splatting for Monocular Dynamic Scene Rendering via Sparse Time-Variant Attribute Modeling (AAAI 2025 · similarity 0.90)
- MoDGS: Dynamic Gaussian Splatting from Casually-captured Monocular Videos with Depth Priors (ICLR 2025 · similarity 0.90)
- 4D Gaussian Splatting in the Wild with Uncertainty-Aware Regularization (NeurIPS 2024 · similarity 0.90)
- BARD-GS: Blur-Aware Reconstruction of Dynamic Scenes via Gaussian Splatting (CVPR 2025 · similarity 0.89)
- 1000+ FPS 4D Gaussian Splatting for Dynamic Scene Rendering (NeurIPS 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.
