본문 바로가기
  • Home

A Study on K-Fantasy Animation Production Pipeline Based on Generative AI

  • Journal of Internet of Things and Convergence
  • Abbr : JKIOTS
  • 2026, 12(1), pp.223~234
  • Publisher : The Korea Internet of Things Society
  • Research Area : Engineering > Computer Science > Internet Information Processing
  • Received : December 24, 2025
  • Accepted : February 19, 2026
  • Published : February 28, 2026

Younghwa Lee 1 Daru Kim 1 O JU Kwon 1 Chaeyoung Moon 1 Gooman Park ORD ID 1 Younghee Choi 1

1서울과학기술대학교

Accredited

ABSTRACT

This study proposes a K-fantasy animation production pipeline utilizing generative AI and verifies its effectiveness through the short animation "Anrisu in Wonderland." Traditional production methods pose high entry barriers for independent creators and small studios due to their high costs and heavy reliance on human labor. To address this, an integrated pipeline combining commercial AI tools like Midjourney, Runway Gen-4, and Higgsfield with LoRA fine-tuning techniques was developed. A strategy of separately fine-tuning character and background models ensured consistency between frames. A dedicated dataset reflecting the structural proportions of traditional Korean architecture and the unique color system of the "Weird Country" was constructed, overcoming existing AI limitations in expressing Korean aesthetics. Expert evaluations confirmed positive results in visual completeness, production consistency, aesthetic expressiveness, and production efficiency. Compared to traditional methods, this approach reduced production time and manpower, proving its potential as a practical alternative for generative AI-based K-content production.

Citation status

* References for papers published after 2024 are currently being built.