CogOmniControl: Reasoning-Driven Controllable Video Generation via Creative Intent Cognition

Recent diffusion models achieve strong photorealism and fluency in video generation, yet remain fragile under abstract, sparse or complex conditions, leading to poor performance in professional production workflows such as storyboard sketches and clay render conditions. Existing video generation mod...