Authoring Comics with AI Assistant

Abstract

This study presents a theory-inspired visual narrative generative system that integrates conceptual principles—comic authoring idioms—with generative and language models to enhance the comic creation process. Instead of relying entirely on generative AI for the final product, this system uses models as components to support parts of the generative process, allowing for a collaborative effort between the author's creativity and AI. The comic-authoring idioms are underlying principles derived from prior human-created image sequences. These idioms serve as guidelines and tools for crafting and refining the storytelling of image sequences. Combined with human creativity, they help convey the narrative clearly and create more engaging experiences for readers. The system translates these conceptual principles into system layers that apply to creating comic content and employ AI models to assist in graphical and overall narrative generation. The system creates comics through sequential decision-making across layers, including panel composition, story tension changes, panel transitions, and narrative elements. Each layer's decisions are based on corresponding narrative goals. The contributions of this comic-generating system are threefold: it supports the integration of machine learning models into the human-AI cooperative comic generation process, it is one of the first systems to deploy abstract narrative theories into AI-driven comic creation, and it provides a tool for customizing comic narratives, offering flexibility and portals for narrative-driven image sequences.