← back to paper
arxiv: 2608.00094 · 2 revisions
Video Models as Native 4D Renderers: World-Grounded Conditioning from Animated Mesh