Designed a production pipeline for the Jeju myth Seolmundae Halmang using generative AI, integrating the visual characteristics of East Asian ink painting with the spatial composition and camera techniques of 3D animation.
Structured the key expressive elements of traditional ink painting—including line, ink tonal gradation, ink diffusion, brushstrokes, negative space, the interplay of presence and absence, and irregular boundaries—and translated them into prompt language interpretable by generative AI.
Conducted a series of prompt experiments using the same scene, progressively incorporating four elements: ① style designation, ② material and expressive characteristics, ③ principles of visual composition, and ④ style suppression constraints. The experiments compared how effectively each approach reproduced the aesthetics of East Asian ink painting.
Integrated controlled image generation with image-to-video production while maintaining consistent character designs, backgrounds, and camera compositions to evaluate the visual consistency of ink-wash aesthetics and their feasibility for animation.
Identified key challenges during the generation process, including unintended photorealism, excessive 3D-rendered textures, stylistic inconsistencies between scenes, and changes in character appearance. Based on these findings, proposed improvements using reference images, standardized style prompts, and negative prompts.

team
Eunsu Ahn(speaker) is a master's student majoring in Film and Media Production at GSAIM, Chung-Ang University. She is particularly interested in using AI to improve the efficiency of filmmaking processes, especially in areas such as visual effects (VFX).
Gilwoo Jeong is a master's student majoring in Film and Media Production at GSAIM, Chung-Ang University. He is interested in the application of AI throughout the filmmaking process, particularly in exploring how AI can improve efficiency in film production.
Hokwon Hwang is a master's student majoring in Art and Technology at GSAIM, Chung-Ang University. His research focuses on integrating art and technology through generative AI-based media content production. He is particularly interested in exploring AI-driven animation production methods and the expressive potential of generative AI.
Hyunji Jang is a master's student majoring in Immersive Animation at GSAIM, Chung-Ang University.She is highly interested in applying AI to animation production, particularly in exploring its potential for creating 3D animation.