I want to create an agentic AI video system that can analyze medical images and an agent can explain a certain part or the entire image.