MSDNet: Efficient 4D Radar Super-Resolution via Multi-Stage Distillation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911157907881984 |
|---|---|
| author | Huang, Minqing Lu, Shouyi Zheng, Boyuan Li, Ziyao Tang, Xiao Zhuo, Guirong |
| author_facet | Huang, Minqing Lu, Shouyi Zheng, Boyuan Li, Ziyao Tang, Xiao Zhuo, Guirong |
| contents | 4D radar super-resolution, which aims to reconstruct sparse and noisy point clouds into dense and geometrically consistent representations, is a foundational problem in autonomous perception. However, existing methods often suffer from high training cost or rely on complex diffusion-based sampling, resulting in high inference latency and poor generalization, making it difficult to balance accuracy and efficiency. To address these limitations, we propose MSDNet, a multi-stage distillation framework that efficiently transfers dense LiDAR priors to 4D radar features to achieve both high reconstruction quality and computational efficiency. The first stage performs reconstruction-guided feature distillation, aligning and densifying the student's features through feature reconstruction. In the second stage, we propose diffusion-guided feature distillation, which treats the stage-one distilled features as a noisy version of the teacher's representations and refines them via a lightweight diffusion network. Furthermore, we introduce a noise adapter that adaptively aligns the noise level of the feature with a predefined diffusion timestep, enabling a more precise denoising. Extensive experiments on the VoD and in-house datasets demonstrate that MSDNet achieves both high-fidelity reconstruction and low-latency inference in the task of 4D radar point cloud super-resolution, and consistently improves performance on downstream tasks. The code will be publicly available upon publication. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_13149 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | MSDNet: Efficient 4D Radar Super-Resolution via Multi-Stage Distillation Huang, Minqing Lu, Shouyi Zheng, Boyuan Li, Ziyao Tang, Xiao Zhuo, Guirong Computer Vision and Pattern Recognition 4D radar super-resolution, which aims to reconstruct sparse and noisy point clouds into dense and geometrically consistent representations, is a foundational problem in autonomous perception. However, existing methods often suffer from high training cost or rely on complex diffusion-based sampling, resulting in high inference latency and poor generalization, making it difficult to balance accuracy and efficiency. To address these limitations, we propose MSDNet, a multi-stage distillation framework that efficiently transfers dense LiDAR priors to 4D radar features to achieve both high reconstruction quality and computational efficiency. The first stage performs reconstruction-guided feature distillation, aligning and densifying the student's features through feature reconstruction. In the second stage, we propose diffusion-guided feature distillation, which treats the stage-one distilled features as a noisy version of the teacher's representations and refines them via a lightweight diffusion network. Furthermore, we introduce a noise adapter that adaptively aligns the noise level of the feature with a predefined diffusion timestep, enabling a more precise denoising. Extensive experiments on the VoD and in-house datasets demonstrate that MSDNet achieves both high-fidelity reconstruction and low-latency inference in the task of 4D radar point cloud super-resolution, and consistently improves performance on downstream tasks. The code will be publicly available upon publication. |
| title | MSDNet: Efficient 4D Radar Super-Resolution via Multi-Stage Distillation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2509.13149 |