Research map / Detection and segmentation / 3D detection and driving
K-Radar: 4D Radar Object Detection for Autonomous Driving in Various Weather Conditions
Dong-Hee Paek, Seung-Hyun Kong, Kevin Tirta Wijaya
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 |
|---|---|---|---|
| use 3D sparse conv backbone with height information (RTNH) instead of 2D dense backbone (RTN) (ablation) | 4DRT baseline NN · K-Radar test (sedan, IoU 0.3) | AP_3D: 40.12 -> 47.44 | Improved |
| use 3D sparse conv backbone with height information instead of 2D dense backbone (ablation) | 4DRT baseline NN · K-Radar test (sedan, IoU 0.3) | AP_BEV: 50.67 -> 58.39 | Improved |
| use sparse 3D convolution on top-10% power RT instead of dense 2D ResNext101 (ablation) | 4DRT baseline NN · K-Radar test (sedan, IoU 0.3) | GPU RAM [MB]: 520 -> 421 | Improved |
| swap RTN backbone from ResNext101 to ResNet50 (ablation) | RTN · K-Radar test (sedan, IoU 0.3) | AP_3D: 40.12 -> 39.86 | No clear effect |
| swap RTN backbone from ResNext101 to ResNet50 (ablation) | RTN · K-Radar test (sedan, IoU 0.3) | GPU RAM [MB]: 520 -> 257 | Improved |
| swap RTN backbone from ResNext101 to ResNet50 (ablation) | RTN · K-Radar test (sedan, IoU 0.3) | AP_BEV: 50.67 -> 49.37 | Hurt |
| use 4D Radar tensor (RTNH) instead of Lidar (PointPillars) | detector · K-Radar test, all weather | AP_3D: 45.4 -> 47.4 | Improved |
| use 4D Radar tensor (RTNH) instead of Lidar (PointPillars) | detector · K-Radar test, all weather | AP_BEV: 49.3 -> 58.4 | Improved |
| use 4D Radar instead of Lidar in sleet | detector · K-Radar test, sleet | AP_3D: 22.7 -> 41.5 | Improved |
| use 4D Radar instead of Lidar in heavy snow | detector · K-Radar test, heavy snow | AP_3D: 29.7 -> 44.5 | Improved |
| use 4D Radar instead of Lidar in fog | detector · K-Radar test, fog | AP_3D: 42.2 -> 52.8 | Improved |
| use 4D Radar instead of Lidar in normal weather | detector · K-Radar test, normal | AP_3D: 52.3 -> 49.9 | Hurt |
Closest papers by meaning
- Towards Robust 3D Object Detection with LiDAR and 4D Radar Fusion in Various Weather Conditions (CVPR 2024 · similarity 0.89)
- RadarOcc: Robust 3D Occupancy Prediction with 4D Imaging Radar (NeurIPS 2024 · similarity 0.89)
- L4DR: LiDAR-4DRadar Fusion for Weather-Robust 3D Object Detection (AAAI 2025 · similarity 0.88)
- V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception (NeurIPS 2025 · similarity 0.87)
- RadarDistill: Boosting Radar-based Object Detection Performance via Knowledge Distillation from LiDAR Features (CVPR 2024 · similarity 0.87)
- SCKD: Semi-Supervised Cross-Modality Knowledge Distillation for 4D Radar Object Detection (AAAI 2025 · similarity 0.86)
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.
