FairyGen: Storied Cartoon Video from a Single Child-Drawn Character
Jiayi Zheng, Xiaodong Cun
2025-06-27
Summary
This paper talks about FairyGen, a system that can create cartoon videos with a story using just one drawing of a character made by a child.
What's the problem?
The problem is that making cartoon videos usually takes a lot of time and effort because artists have to design characters, backgrounds, and movement separately, which makes it hard to quickly turn a simple drawing into a full video.
What's the solution?
The researchers designed FairyGen to separate the character from the background and use large language models to create storyboards that plan the story. They use special techniques to keep the style consistent throughout the video and apply advanced models to generate smooth movements, turning that single drawing into a full cartoon story.
Why it matters?
This matters because it makes it much easier and faster for anyone, especially kids, to bring their drawings to life in videos, encouraging creativity and storytelling with the help of AI.
Abstract
FairyGen generates story-driven cartoon videos from a single drawing by disentangling character modeling and background styling, employing MLLM for storyboards, style propagation for consistency, and MMDiT-based diffusion models for motion.