DyRAD: Radar Novel View Synthesis for Dynamic Driving Scenes
DyRAD renders moving driving scenes as full range-azimuth-Doppler radar, using Doppler rather than leaving it on the floor.
The paper models static background reflectors separately from motion-tracked dynamic point reflectors, then projects their velocities into the radar line of sight. It also uses an analytic point-spread function from the radar processing chain, so sensor blur is not mistaken for scene geometry. The authors say that separation lets the same reconstructed scene render under different radar specifications without refitting. On RADIal, DyRAD recovered detections for 90.7% of reference-detected objects, versus 26.9% for the strongest baseline. HF Daily Papers' note
The paper models static background reflectors separately from motion-tracked dynamic point reflectors, then projects their velocities into the radar line of sight. It also uses an analytic point-spread function from the radar processing chain, so sensor blur is not mistaken for scene geometry. The authors say that separation lets the same reconstructed scene render under different radar specifications without refitting. On RADIal, DyRAD recovered detections for 90.7% of reference-detected objects, versus 26.9% for the strongest baseline. HF Daily Papers' note
score 4