1 paper · 1 filter
Chuanzhi Xu, Huiqi Liang, Bang Shi +7
Long video generation requires high-fidelity visual synthesis, coherent narrative organization, shot-level structure, and explicit user control. Existing text-to-video methods typi…