We repurpose video diffusion models to reconstruct 4D geometry (4D pointmaps) from monocular video.
Mar 16, 2025
