![[2602.11554] HyperDet: 3D Object Detection with Hyper 4D Radar Point Clouds](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.11554] HyperDet: 3D Object Detection with Hyper 4D Radar Point Clouds
Summary
The paper presents HyperDet, a novel framework for 3D object detection using hyper 4D radar point clouds, addressing limitations of traditional radar systems compared to LiDAR.
Why It Matters
As the demand for robust and cost-effective 3D object detection grows, HyperDet offers a significant advancement in radar technology, potentially transforming applications in robotics and autonomous vehicles by improving detection accuracy and reliability in various conditions.
Key Takeaways
- HyperDet enhances radar-based 3D detection by aggregating data from multiple radar sources.
- The framework employs a self-consistency check to improve data reliability and suppress noise.
- Integration of radar and LiDAR supervision helps in refining object structures.
- HyperDet narrows the performance gap between radar and LiDAR systems without requiring architectural changes.
- The approach demonstrates improved detection accuracy in real-world scenarios like MAN TruckScenes.
Computer Science > Robotics arXiv:2602.11554 (cs) [Submitted on 12 Feb 2026 (v1), last revised 13 Feb 2026 (this version, v2)] Title:HyperDet: 3D Object Detection with Hyper 4D Radar Point Clouds Authors:Yichun Xiao, Runwei Guan, Fangqiang Ding View a PDF of the paper titled HyperDet: 3D Object Detection with Hyper 4D Radar Point Clouds, by Yichun Xiao and 2 other authors View PDF HTML (experimental) Abstract:4D mmWave radar provides weather-robust, velocity-aware measurements and is more cost-effective than LiDAR. However, radar-only 3D detection still trails LiDAR-based systems because radar point clouds are sparse, irregular, and often corrupted by multipath noise, yielding weak and unstable geometry. We present HyperDet, a detector-agnostic radar-only 3D detection framework that constructs a task-aware hyper 4D radar point cloud for standard LiDAR-oriented detectors. HyperDet aggregates returns from multiple surround-view 4D radars over consecutive frames to improve coverage and density, then applies geometry-aware cross-sensor consensus validation with a lightweight self-consistency check outside overlap regions to suppress inconsistent returns. It further integrates a foreground-focused diffusion module with training-time mixed radar-LiDAR supervision to densify object structures while lifting radar attributes (e.g., Doppler, RCS); the model is distilled into a consistency model for single-step inference. On MAN TruckScenes, HyperDet consistently improves over raw ra...