Efficiently Reconstructing Dynamic Scenes One D4RT at a Time
IntermediateChuhan Zhang, Guillaume Le Moing et al.Dec 9arXiv
D4RT is a new AI model that turns regular videos into moving 3D scenes (4D) quickly and accurately.
#D4RT#dynamic 4D reconstruction#query-based decoding