Research map / 3D vision / Gaussian splatting
SGS-SLAM: Semantic Gaussian Splatting For Neural Dense SLAM
Mingrui Li, Shuhong Liu, Heng Zhou, Guohao Zhu, Na Cheng, Tianchen Deng, Hongyu 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 semantic map supervision to multi-channel optimization (ablation) | SGS-SLAM · ScanNet scene0000_00 | ATE RMSE (cm): 13.81 -> 11.26 | Improved |
| add depth supervision (ablation) | SGS-SLAM · ScanNet scene0000_00 | Depth L1 (cm): 47.66 -> 6.18 | Improved |
| add color image supervision (ablation) | SGS-SLAM · ScanNet scene0000_00 | ATE RMSE (cm): 24.59 -> 11.26 | Improved |
| add silhouette threshold in tracking (ablation) | SGS-SLAM · ScanNet scene0000_00 | ATE RMSE (cm): 357.48 -> 11.26 | Improved |
| add geometric overlap threshold in keyframe selection (ablation) | SGS-SLAM · ScanNet scene0000 | ATE RMSE (cm): 15.55 -> 11.26 | Improved |
| add semantic threshold in keyframe selection (ablation) | SGS-SLAM · ScanNet scene0000 | PSNR (dB): 17.84 -> 19.47 | Improved |
| add uncertainty weighting of keyframes (ablation) | SGS-SLAM · ScanNet scene0000 | PSNR (dB): 18.72 -> 19.47 | Improved |
| use Gaussian splatting with semantic channel instead of NeRF-based semantic SLAM | SGS-SLAM · Replica (4 scenes) | avg mIoU (%): 87.41 -> 92.72 | Improved |
| add semantic channel and keyframe selection to Gaussian SLAM (vs SplaTAM) | SGS-SLAM · Replica (8 scenes avg) | Depth L1 (cm): 0.525 -> 0.356 | Improved |
| SGS-SLAM vs SplaTAM on runtime | SGS-SLAM · Replica (8 scenes avg) | SLAM FPS: 2.26 -> 2.11 | Hurt |
| SGS-SLAM vs ESLAM on ScanNet tracking | SGS-SLAM · ScanNet (6 scenes) | ATE RMSE avg (cm): 7.88 -> 9.87 | Hurt |
| Gaussian representation vs ESLAM on large-motion ScanNet++ | SGS-SLAM · ScanNet++ (2 scenes) | ATE RMSE avg (cm): 170.06 -> 1.62 | Improved |
Later papers that compared with this method
2 later accepted papers report SGS-SLAM as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- UniGS: Unified Geometry-Aware Gaussian Splatting for Multimodal Rendering (NeurIPS 2026 · better on 3 of 3 shared results)
- Open-Set Semantic Gaussian Splatting SLAM with Expandable Representation (ICLR 2026 · better on 1 of 1 shared result)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 4 of 5 on most shared results.
- SplaTAM — SplaTAM: Splat Track & Map 3D Gaussians for Dense RGB-D SLAM (CVPR 2024) · better on 5 of 6 shared results
- Co-SLAM — Co-SLAM: Joint Coordinate and Sparse Parametric Encodings for Neural Real-Time SLAM (CVPR 2023) · better on 1 of 3 shared results
- ESLAM — ESLAM: Efficient Dense SLAM System Based on Hybrid Representation of Signed Distance Fields (CVPR 2023) · better on 2 of 3 shared results
- SNI-SLAM — SNI-SLAM: Semantic Neural Implicit SLAM (CVPR 2024) · better on 1 of 1 shared result
- NICE-SLAM — NICE-SLAM: Neural Implicit Scalable Encoding for SLAM (CVPR 2022) · 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
- GS-SLAM: Dense Visual SLAM with 3D Gaussian Splatting (CVPR 2024 · similarity 0.92)
- SNI-SLAM: Semantic Neural Implicit SLAM (CVPR 2024 · similarity 0.90)
- VTGaussian-SLAM: RGBD SLAM for Large Scale Scenes with Splatting View-Tied 3D Gaussians (ICML 2025 · similarity 0.89)
- Gaussian Splatting SLAM (CVPR 2024 · similarity 0.89)
- DG-SLAM: Robust Dynamic Gaussian Splatting SLAM with Hybrid Pose Optimization (NeurIPS 2024 · similarity 0.89)
- WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments (CVPR 2025 · similarity 0.88)
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.
