Text-to-video synthesis

This study investigates the capabilities of video diffusion models in generating human motion from text prompts, revealing their strengths in common motions and limitations in rare or complex movements.