Research map / 3D vision / Gaussian splatting
RadarSplat: Radar Gaussian Splatting for High-Fidelity Data Synthesis and 3D Reconstruction of Autonomous Driving Scenes
Pou-Chun Kung, Skanda Harisha, Ram Vasudevan, Aline Eid, Katherine A. Skinner
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 noise probability term to Gaussian power return (ablation) | RadarSplat · Boreas image synthesis | PSNR: 23.52 -> 26.06 | Improved |
| add noise probability term to Gaussian power return (ablation) | RadarSplat · Boreas image synthesis | SSIM: 0.23 -> 0.51 | Improved |
| add multipath modeling (ablation) | RadarSplat · Boreas image synthesis | PSNR: 25.95 -> 26.06 | Improved |
| add multipath modeling in urban scenes (ablation) | RadarSplat · Boreas urban scenes | PSNR: 25.25 -> 26.06 | Improved |
| add multipath modeling in natural scenes (ablation) | RadarSplat · Boreas natural scenes | PSNR: 27.27 -> 27.27 | No clear effect |
| add occupancy map supervision (ablation) | RadarSplat · Boreas scene reconstruction | Accuracy: 0.30 -> 0.91 | Improved |
| add spectral leakage modeling (ablation) | RadarSplat · Boreas scene reconstruction | RMSE: 2.05 -> 1.81 | Improved |
| use proposed denoised occupancy map instead of Radar Fields map as initial map | Initial occupancy map · Boreas scene reconstruction | RMSE: 3.40 -> 1.81 | Improved |
| use proposed occupancy map in Radar Fields (ablation) | Radar Fields · Boreas scene reconstruction | Accuracy: 0.59 -> 0.72 | Improved |
| use proposed occupancy map in RadarSplat (ablation) | RadarSplat · Boreas scene reconstruction | RMSE: 1.83 -> 1.81 | Improved |
| replace Radar Fields NeRF with RadarSplat | Radar Fields · Boreas image synthesis | PSNR: 22.66 -> 26.06 | Improved |
| replace Radar Fields NeRF with RadarSplat | Radar Fields · Boreas scene reconstruction | RMSE: 3.03 -> 1.81 | Improved |
Later papers that compared with this method
1 later accepted paper reports RadarSplat as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- 3D Point Splatting for mmWave Radar Novel View Synthesis (NeurIPS 2026 · better on 8 of 8 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
- SplatAD: Real-Time Lidar and Camera Rendering with 3D Gaussian Splatting for Autonomous Driving (CVPR 2025 · similarity 0.88)
- DrivingGaussian: Composite Gaussian Splatting for Surrounding Dynamic Autonomous Driving Scenes (CVPR 2024 · similarity 0.87)
- DeSiRe-GS: 4D Street Gaussians for Static-Dynamic Decomposition and Surface Reconstruction for Urban Driving Scenes (CVPR 2025 · similarity 0.86)
- D$^2$GS: Dense Depth Regularization for LiDAR-free Urban Scene Reconstruction (NeurIPS 2025 · similarity 0.86)
- EVolSplat: Efficient Volume-based Gaussian Splatting for Urban View Synthesis (CVPR 2025 · similarity 0.85)
- Horizon-GS: Unified 3D Gaussian Splatting for Large-Scale Aerial-to-Ground Scenes (CVPR 2025 · 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.
