Research map / 3D vision / Gaussian splatting
Underwater Visual SLAM with Depth Uncertainty and Medium Modeling
Rui Liu, Sheng Fan, Wenguan Wang, Yi Yang
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 underwater medium modeling (neural scattering field) to 3DGS mapping (ablation) | DUV-SLAM · TartanAir-Hard | PSNR: 24.08 -> 27.57 | Improved |
| use separate attenuation and backscatter coefficients instead of a single shared one (ablation) | DUV-SLAM · TartanAir-Hard | PSNR: 26.93 -> 27.57 | Improved |
| use separate attenuation and backscatter coefficients (ablation) | DUV-SLAM · TartanAir-Hard | LPIPS/SSIM: SSIM 0.69 -> 0.67; LPIPS 0.29 -> 0.25 | Mixed |
| add Mahalanobis depth loss weighted by tracking uncertainty (LMD) (ablation) | DUV-SLAM · TartanAir-Hard | PSNR: 25.60 -> 27.57 | Improved |
| add KL distribution loss LKL (ablation) | DUV-SLAM · TartanAir-Hard | PSNR: 26.70 -> 27.57 | Improved |
| add all depth losses (LMD, LRGB-C, LKL) vs none (ablation) | DUV-SLAM · FLSea Canyons-Fla | PSNR: 23.77 -> 25.71 | Improved |
| add medium modeling to MonoGS and Photo-SLAM trackers (ablation) | MonoGS, Photo-SLAM · FLSea Canyons-TC | PSNR: Photo-SLAM 24.16 -> 26.07 | Improved |
| DUV-SLAM vs DROID-SLAM on tracking accuracy | DUV-SLAM · TartanAir-Ocean Easy | ATE: 1.15 -> 1.26 | Hurt |
| DUV-SLAM uncertainty vs UncLe-SLAM uncertainty | DUV-SLAM · uncertainty AUSE average | AUSE: 0.25 -> 0.21 | Improved |
| DUV-SLAM on long sequences | DUV-SLAM · MIMIR-UW Ocean Floor t1-l | ATE (m): 15.59 | Hurt |
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.
- Photo-SLAM — Photo-SLAM: Real-time Simultaneous Localization and Photorealistic Mapping for Monocular Stereo and RGB-D Cameras (CVPR 2024) · better on 5 of 6 shared results
- DROID-SLAM — DROID-SLAM: Deep Visual SLAM for Monocular, Stereo, and RGB-D Cameras (NeurIPS 2021) · better on 3 of 4 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
- WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments (CVPR 2025 · similarity 0.86)
- Atlantis: Enabling Underwater Depth Estimation with Stable Diffusion (CVPR 2024 · similarity 0.86)
- WaterDiffusion: Learning a Prior-involved Unrolling Diffusion for Joint Underwater Saliency Detection and Visual Restoration (AAAI 2025 · similarity 0.85)
- 4DTAM: Non-Rigid Tracking and Mapping via Dynamic Surface Gaussians (CVPR 2025 · similarity 0.85)
- DG-SLAM: Robust Dynamic Gaussian Splatting SLAM with Hybrid Pose Optimization (NeurIPS 2024 · similarity 0.85)
- GS-SLAM: Dense Visual SLAM with 3D Gaussian Splatting (CVPR 2024 · similarity 0.85)
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.
