milliFlow proposes a deep learning approach to estimate scene flow for mmWave radar point clouds, enhancing human motion sensing tasks.
mmWave radar-based scene flow estimation enhances human motion sensing with deep learning.
Diffusion-based refinement improves scene flow estimation robustness and reliability.
We reframe scene flow estimation as the task of fitting a continuous Partial Differential Equation (PDE) that describes the motion of an entire observation sequence, enabling high quality scene flow estimation via self-supervision.