Tag
This arXiv paper presents the Dream Scene Visualiser (DSV), a system that converts written dream descriptions into a coherent four-panel visual sequence using an LLM for narrative splitting and text-to-image models with feedback for visual and semantic coherence.