Research map / 3D vision / Gaussian splatting
ODGS-SLAM: Omnidirectional Gaussian Splatting SLAM
Stefan Spiss, Joey Hieronimy, Marcel Ritter, Matthias Harders
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 |
|---|---|---|---|
| replace perspective 3DGS-SLAM (MonoGS) by omnidirectional ODGS-SLAM | 3DGS-SLAM with RGB input · Synthetic sequences, RGB (average) | RMSE ATE (m): 0.663 -> 0.068 | Improved |
| replace MonoGS by ODGS-SLAM with RGBD input | 3DGS-SLAM with RGBD input · Synthetic sequences, RGBD (average) | RMSE ATE (m): 0.579 -> 0.029 | Improved |
| use ODGS-SLAM instead of feature-based omnidirectional PMSP | omnidirectional SLAM · Insta360 Pro sequences, RGB (average) | RMSE ATE (m): 0.122 -> 0.031 | Improved |
| use ODGS-SLAM instead of PMSP on one long synthetic exploration sequence (Ex2, RGB) | omnidirectional SLAM · Synthetic Ex2, RGB | RMSE ATE (m): 0.138 (PMSP) -> 0.355 | Hurt |
| add omnidirectional rendering with RGB-only input outdoors | ODGS-SLAM RGB · Outdoor synthetic Ex-O, RGB | RMSE ATE (m): 2.365 (PMSP) -> 36.070 | Hurt |
| add latitude weighting to the photometric and depth losses (ablation) | ODGS-SLAM RGBD · Synthetic RGBD, Ex1 first 750 frames | RMSE ATE (m): 0.030 -> 0.007 | Improved |
| add latitude weighting (ablation) | ODGS-SLAM RGBD · Synthetic RGBD, Ex1 first 750 frames | Track time (s): 1.412 -> 1.532 | Hurt |
| add large depth masking (ablation) | ODGS-SLAM RGBD · Synthetic RGBD, Ex1 first 750 frames | RMSE ATE (m): 0.028 -> 0.007 | Improved |
| add constant motion assumption for pose initialization (ablation) | ODGS-SLAM RGBD · Synthetic RGBD, Ex1 first 750 frames | RMSE ATE (m): 0.006 -> 0.007 | No clear effect |
| add graph-based key frame removal (ablation) | ODGS-SLAM RGBD · Synthetic RGBD, Ex1 first 750 frames | BE GPU memory (MB): 8857.2 -> 7906.8 | Improved |
| add graph-based key frame removal (ablation) | ODGS-SLAM RGBD · Synthetic RGBD, Ex1 first 750 frames | PSNR (dB): 28.99 -> 27.84 | Hurt |
| add latitude weighting (ablation) | ODGS-SLAM RGBD · Synthetic RGBD, Ex1 first 750 frames | PSNR (dB): 28.33 -> 27.84 | Hurt |
Closest papers by meaning
- Gaussian Splatting SLAM (CVPR 2024 · similarity 0.92)
- ODGS: 3D Scene Reconstruction from Omnidirectional Images with 3D Gaussian Splattings (NeurIPS 2024 · similarity 0.91)
- DG-SLAM: Robust Dynamic Gaussian Splatting SLAM with Hybrid Pose Optimization (NeurIPS 2024 · similarity 0.91)
- GS-SLAM: Dense Visual SLAM with 3D Gaussian Splatting (CVPR 2024 · similarity 0.90)
- SplaTAM: Splat Track & Map 3D Gaussians for Dense RGB-D SLAM (CVPR 2024 · similarity 0.89)
- WildGS-SLAM: Monocular Gaussian Splatting SLAM in Dynamic Environments (CVPR 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.
