ai-video VIRTUE: New AI Model Enables Precise Visual Interactions Researchers introduce VIRTUE, a visual-interactive embedding model that can understand specific image regions through user prompts, advancing capabilities for synthetic media.