Research map / Point clouds / Segmentation and detection
DoppDrive: Doppler-Driven Temporal Aggregation for Improved Radar Object Detection
Yuval Haitman, Oded Bialer
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 Doppler-driven aggregation versus standard ego-compensated aggregation | SMURF · aiMotive | AP: 81.7 -> 89.1 | Improved |
| add radial Doppler compensation to aggregation (ablation) | DoppDrive with SMURF · aiMotive | AP: 81.4 -> 85.2 | Improved |
| add variable per-point aggregation duration (ablation) | DoppDrive with SMURF · aiMotive | AP: 85.2 -> 89.1 | Improved |
| estimate ego velocity from radar point cloud instead of GPS-INS (ablation) | DoppDrive with SMURF · aiMotive | AP: 89.1 -> 87.7 | Hurt |
| remove ego-speed Doppler component from point features (ablation) | SMURF with no aggregation · aiMotive | AP: 66.3 -> 73.1 | Improved |
| increase tolerated offset D from 2 m to 5 m (ablation) | DoppDrive · LRR-Sim | AP: 93.3 -> 89.6 | Hurt |
| decrease tolerated offset D from 2 m to 1 m (ablation) | DoppDrive · LRR-Sim | AP: 93.3 -> 91.2 | Hurt |
| apply DoppDrive at long range beyond 175 m | SMURF · LRR-Sim range > 175 m | AP: 63.1 -> 69.2 | Improved |
| use the full multi-modal heading histogram instead of Laplace approximation for g(theta) (ablation) | DoppDrive with SMURF · aiMotive | AP: +0.4 AP | No clear effect |
| add DoppDrive aggregation step to inference pipeline | SMURF · V100 runtime | total time (ms): 66.4 -> 69.5 | No clear effect |
Closest papers by meaning
- Doppler-Aware LiDAR-RADAR Fusion for Weather-Robust 3D Detection (ICCV 2025 · similarity 0.86)
- HGSFusion: Radar-Camera Fusion with Hybrid Generation and Synchronization for 3D Object Detection (AAAI 2025 · similarity 0.84)
- RadarOcc: Robust 3D Occupancy Prediction with 4D Imaging Radar (NeurIPS 2024 · similarity 0.83)
- AS-Det: Active Sampling for Adaptive 3D Object Detection in Point Clouds (AAAI 2025 · similarity 0.83)
- RadarMOSEVE: A Spatial-Temporal Transformer Network for Radar-Only Moving Object Segmentation and Ego-Velocity Estimation (AAAI 2024 · similarity 0.83)
- Instantaneous Perception of Moving Objects in 3D (CVPR 2024 · similarity 0.82)
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.
