🧪 Test?View on arXiv
GRADE: Single-Frame Generative Radar Depth Estimation Under Visual Degradation
Not provided in the content
depth estimationradar technologygenerative modelsenvironmental robustness
2609.10756
Builder Relevance
4h ago80%
Abstract
GRADE is a method that improves depth perception in challenging visual conditions using mmWave radar.
Reality Card
Core Claim
GRADE achieves a mean absolute error (MAE) of 0.303 m in clear scenes and 0.313 m under smoke, outperforming existing baselines.
Method / Result
Trained and evaluated on ~95K frames across 12 buildings, achieving an MAE of 0.303 m in clear scenes.
Limitations
The method's performance may vary significantly under different environmental conditions not covered in training.
Paper to code
Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.
No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.