# TransRAD: Retentive Vision Transformer for Enhanced Radar Object Detection

arXiv: https://arxiv.org/abs/2501.17977 · HTML: https://mlsanity.com/paper/2501.17977/

## Key facts

- Cited by 8 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 29 Jan 2025.

## Built on by (most cited first, 8 of 8)

- [CalibRefine: Deep Learning-Based Online Automatic Targetless LiDAR-Camera Calibration with Iterative and Attention-Driven Post-Refinement](https://mlsanity.com/paper/2502.17648/) (2025)
- [RADE-Net: Robust Attention Network for Radar-Only Object Detection in Adverse Weather](https://mlsanity.com/paper/2602.19994/) (2026)
- [A Survey of Retentive Network](https://mlsanity.com/paper/2506.06708/) (2025)
- [mRadNet: A Compact Radar Object Detector with MetaFormer](https://mlsanity.com/paper/2509.16223/) (2025)
- [Radar-Camera Fused Multi-Object Tracking: Online Calibration and Common Feature](https://mlsanity.com/paper/2510.20794/) (2025)
- [Synthetic FMCW Radar Range Azimuth Maps Augmentation with Generative Diffusion Model](https://mlsanity.com/paper/2601.06228/) (2026)
- [Leveraging Transformer Decoder for Automotive Radar Object Detection](https://mlsanity.com/paper/2601.13386/) (2026)
- [Revisiting Radar Perception With Spectral Point Clouds](https://mlsanity.com/paper/2604.08282/) (2026)

## Abstract (arXiv)

Despite significant advancements in environment perception capabilities for autonomous driving and intelligent robotics, cameras and LiDARs remain notoriously unreliable in low-light conditions and adverse weather, which limits their effectiveness. Radar serves as a reliable and low-cost sensor that can effectively complement these limitations. However, radar-based object detection has been underexplored due to the inherent weaknesses of radar data, such as low resolution, high noise, and lack of visual information. In this paper, we present TransRAD, a novel 3D radar object detection model designed to address these challenges by leveraging the Retentive Vision Transformer (RMT) to more effectively learn features from information-dense radar Range-Azimuth-Doppler (RAD) data. Our approach leverages the Retentive Manhattan Self-Attention (MaSA) mechanism provided by RMT to incorporate explicit spatial priors, thereby enabling more accurate alignment with the spatial saliency characteristics of radar targets in RAD data and achieving precise 3D radar detection across Range-Azimuth-Doppler dimensions. Furthermore, we propose Location-Aware NMS to effectively mitigate the common issue of duplicate bounding boxes in deep radar object detection. The experimental results demonstrate that TransRAD outperforms state-of-the-art methods in both 2D and 3D radar detection tasks, achieving higher accuracy, faster inference speed, and reduced computational complexity. Code is available at https://github.com/radar-lab/TransRAD