Research map / Detection and segmentation / Tracking
RadarNet: Exploiting Radar for Robust Perception of Dynamic Objects
Bin Yang, Runsheng Guo, Ming Liang, Sergio Casas, Raquel Urtasun
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 Radar early fusion to LiDAR detector (ablation) | RadarNet · nuScenes val, motorcycles | AP@2m: 53.7 -> +1.9 | Improved |
| add Radar early fusion to LiDAR detector (ablation) | RadarNet · nuScenes val, cars | AP@2m: 87.6 -> +0.3 | Improved |
| add Radar early fusion (ablation) | RadarNet · nuScenes val, cars | AVE: -2% | No clear effect |
| add Radar early fusion (ablation) | RadarNet · DenseRadar val, 70-100m | Vehicles AP: 77.5 -> +0.8 | Improved |
| add attention-based late fusion after early fusion (ablation) | RadarNet · nuScenes val, cars | AVE: -14% | Improved |
| add attention-based late fusion after early fusion (ablation) | RadarNet · nuScenes val, motorcycles | AVE: -15% | Improved |
| add attention-based late fusion after early fusion (ablation) | RadarNet · DenseRadar val | ADVE: -3% -> -19% | Improved |
| use attention-based late fusion instead of heuristic late fusion (ablation) | RadarNet · nuScenes val, cars | AVE: -9% -> -14% | Improved |
| use attention-based late fusion instead of heuristic late fusion (ablation) | RadarNet · DenseRadar val | ADVE: -6% -> -19% | Improved |
| add heuristic late fusion after early fusion (ablation) | RadarNet · DenseRadar val | ADVE: -3% -> -6% | Improved |
| fuse LiDAR and Radar versus best prior LiDAR (CBGS) | RadarNet · nuScenes val, cars | AP: 82.3 -> 84.5 | Improved |
| fuse LiDAR and Radar versus best prior LiDAR (CBGS) | RadarNet · nuScenes val, motorcycles | AVE: 0.339 -> 0.269 | Improved |
Closest papers by meaning
- SparseRadNet: Sparse Perception Neural Network on Subsampled Radar Data (ECCV 2024 · similarity 0.85)
- Exploiting Temporal Relations on Radar Perception for Autonomous Driving (CVPR 2022 · similarity 0.85)
- RETR: Multi-View Radar Detection Transformer for Indoor Perception (NeurIPS 2024 · similarity 0.84)
- Learning Structurally Consistent Representations for Multi-View Radar Semantic Segmentation (ECCV 2026 · similarity 0.84)
- V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception (NeurIPS 2025 · similarity 0.82)
- SIRA: Scalable Inter-frame Relation and Association for Radar Perception (CVPR 2024 · similarity 0.81)
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.
