Research map / 3D vision / Gaussian splatting
Robust and Efficient Monocular 3D Gaussian SLAM for Kilometer-Scale Outdoor Scenes
Sicheng Yu, Dongxu Shen, Beizhen ZHAO, Ding Guanzhi, Hao Wang
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 foundation model activation to tracking (ablation) | KiloGS-SLAM · KITTI-07 | ATE RMSE (m): 2.32 -> 1.06 | Improved |
| use dual-modal (Essential + PnP) switching instead of Essential only (ablation) | KiloGS-SLAM · KITTI-07 | ATE RMSE (m): 2.15 -> 1.06 | Improved |
| use dual-modal switching instead of PnP only (ablation) | KiloGS-SLAM · KITTI-07 | ATE RMSE (m): 3.64 -> 1.06 | Improved |
| add dynamic filtering via sparse optical flow (ablation) | KiloGS-SLAM · KITTI-07 | ATE RMSE (m): 1.81 -> 1.06 | Improved |
| add loop closure (ablation) | KiloGS-SLAM · KITTI (avg ATE) | ATE RMSE (m): 4.26 -> 3.46 | Improved |
| use DoG instead of Canny for initialization (ablation) | KiloGS-SLAM · KITTI-02 | PSNR: 21.05 -> 21.77 | Improved |
| use DoG instead of LoG for initialization (ablation) | KiloGS-SLAM · KITTI-02 | PSNR: 21.62 -> 21.77 | Improved |
| add depth gradient term to probabilistic initialization (ablation) | KiloGS-SLAM · KITTI-02 | PSNR: 21.56 -> 21.77 | Improved |
| use probabilistic DoG initialization instead of random sampling (ablation) | KiloGS-SLAM · KITTI-02 | #Gaussians (M) / VRAM: 365M -> 135M Gaussians, 44 -> 19 GB, PSNR 21.84 -> 21.77 | Mixed |
| add NMS to initialization (ablation) | KiloGS-SLAM · KITTI-02 | #Gaussians (M): 202 -> 135 (PSNR 21.79 -> 21.77) | Improved |
| add multi-view densification (ablation) | KiloGS-SLAM · KITTI-02 | PSNR: 21.16 -> 21.77 | Improved |
| add multi-view pruning (ablation) | KiloGS-SLAM · KITTI-02 | #Gaussians (M): 166 -> 135 (PSNR 21.71 -> 21.77) | Improved |
Closest papers by meaning
- Gaussian Splatting SLAM (CVPR 2024 · similarity 0.93)
- DG-SLAM: Robust Dynamic Gaussian Splatting SLAM with Hybrid Pose Optimization (NeurIPS 2024 · similarity 0.92)
- GS-SLAM: Dense Visual SLAM with 3D Gaussian Splatting (CVPR 2024 · similarity 0.91)
- WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments (CVPR 2025 · similarity 0.90)
- VTGaussian-SLAM: RGBD SLAM for Large Scale Scenes with Splatting View-Tied 3D Gaussians (ICML 2025 · similarity 0.89)
- ProDyG: Progressive Dynamic Scene Reconstruction via Gaussian Splatting from Monocular Videos (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.
