Translated data: The researchers from the University of Science and Technology of China, Microsoft, and others have proposed a video generation model called DragNUWA, which is based on open-domain diffusion. By integrating text, image, and trajectory controls, DragNUWA enables fine-grained control over video content. Experiments have demonstrated that DragNUWA boasts exceptional performance, reliably controlling complex movements and generating stable, coherent videos.