Research map / Architectures and efficiency / Quantization
AdaRadar: Rate Adaptive Spectral Compression for Radar-based Perception
Jinho Park, Se Young Chun, Mingoo Seok
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 index-value top-K pruning with adaptive spectral (DCT) pruning plus quantization (ablation) | FFTRadNet on RADIal · RADIal test (detection) | F1: 75.91 -> 95.13 | Improved |
| replace index-value compression with adaptive spectral compression (ablation) | TMVA-Net on CARRADA · CARRADA test (segmentation) | mIoU: 38.96 -> 54.03 | Improved |
| apply 101x adaptive compression versus uncompressed input | FFTRadNet on RADIal · RADIal test (freeway segmentation) | mIoU: 75.97 -> 79.34 | Improved |
| apply 101x adaptive compression versus uncompressed input | FFTRadNet on RADIal · RADIal test (detection) | F1: 96.58 -> 95.13 | Hurt |
| apply 30x adaptive compression versus uncompressed input | Radatron baseline · Radatron test | AP75: 44.16 -> 49.07 | Improved |
| replace index-value with adaptive spectral compression (ablation) | Radatron baseline · Radatron test | AP50: 80.44 -> 83.69 | Improved |
| replace index-value with adaptive spectral compression (ablation) | RADIal detection · RADIal test (detection) | precision: 97.55 -> 96.25 | Hurt |
| spectral-domain pruning instead of spatial top-K pruning (static) (ablation) | FFTRadNet on RADIal · RADIal test (static pruning, Float32) | F1 roll-off per decade: 46.6%/dec -> 23.1%/dec | Improved |
| prune at ratio 5 without quantization (qualitative) | FFTRadNet on RADIal · RADIal qualitative | detection quality | Improved |
Closest papers by meaning
- Rethinking Radar 3D Perception: Representation Learning from Raw Spectra (NeurIPS 2026 · similarity 0.88)
- V2X-Radar: A Multi-modal Dataset with 4D Radar for Cooperative Perception (NeurIPS 2025 · similarity 0.85)
- Bootstrapping Autonomous Driving Radars with Self-Supervised Learning (CVPR 2024 · similarity 0.84)
- RadarOcc: Robust 3D Occupancy Prediction with 4D Imaging Radar (NeurIPS 2024 · similarity 0.84)
- RadSimReal: Bridging the Gap Between Synthetic and Real Data in Radar Object Detection With Simulation (CVPR 2024 · similarity 0.83)
- AdaPKC: PeakConv with Adaptive Peak Receptive Field for Radar Semantic Segmentation (NeurIPS 2024 · similarity 0.83)
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.
