Any4D: Open-Prompt 4D Generation from Natural Language and Images
arXiv:2511.18746v2 Announce Type: replace
Abstract: While video-generation-based embodied world models have gained increasing attention, their reliance on large-scale embodied interaction data remains a key bottleneck. The scarcity, difficulty of coll…