US20250322176
2025-10-16
Physics
G06F40/40
An AI-based illustrated story generation system is designed to create personalized stories with stylized illustrations. The system uses user-provided photos to generate character profiles, which are then used in conjunction with narrative prompts to produce illustrated stories. The character profiles are created by a character description model that interprets the photos, followed by an image generating model that produces stylized images. The story text is generated by a story text model and accompanied by illustrations created page by page, ensuring visual consistency and character development throughout the narrative.
Illustrated stories combine narratives with visual elements to enhance storytelling, offering a rich experience that engages both imagination and senses. Current generative models can create detailed images but struggle with maintaining consistent character representations across a sequence of illustrations or showing believable character development. There's a need for AI-based systems that can generate coherent story texts and illustrations while allowing personalization and user-guided content creation.
The system comprises a data processing setup that includes a processor and memory, executing instructions to perform several tasks. It receives narrative prompts defining story parameters and identifies character profiles for inclusion in the story. These inputs are processed by a story text generator model to produce multi-page story texts. Each page's text is then used as input for a page illustration generator model, which creates corresponding illustrations. These elements are saved together as part of an illustrated story file.
The system addresses challenges of current models by offering interactive and personalized storytelling experiences with consistent character illustrations. It utilizes AI technologies like Large Language Models (LLMs) to create immersive visual storytelling. The architecture includes modules for maintaining visual identity consistency, modeling character development, and adapting storylines interactively, resulting in personalized storybooks responsive to user inputs.
The interactive story generation service operates as a cloud-based service involving client devices communicating via networks like LANs or WANs. It utilizes servers for computational resources necessary for the service's implementation, including data storage for managing the interactive storytelling process. This setup supports various communication protocols, ensuring robust data transmission and interaction capabilities for creating dynamic illustrated stories.