ScholarWorks@숙명여자대학교

초록

The rapid advancement of generative artificial intelligence (AI) has fundamentally transformed traditional workflows and production stages in animation. In particular, the development of text-, image-, and video-generation technologies has enabled scenes and motions to be generated and reviewed instantly even at the rough animation stage, calling for a redefinition of the role and conceptual boundaries of this phase. While rough animation has traditionally functioned as an intermediate step for verifying timing and staging, AI-based production environments have shifted the overall process toward a non-linear and iterative exploratory structure. This study aims to analyze how the role of rough animation is expanded and reconfigured through experimental animation production using generative AI. To this end, AI-assisted rough animation practices are categorized into three types: idea generation– oriented exploration, motion transfer–oriented exploration, and visual anchor– based exploration. Under identical character and scene conditions, the production characteristics and creator decision-making structures of each exploration type are comparatively examined. The results indicate that each type exhibits distinct differences in motion generation mechanisms, levels of control, and the timing of creative decision-making, demonstrating that rough animation functions not as a single fixed stage but as a combination of multi-layered exploratory strategies selectively employed according to production objectives. Based on these findings, this study reconceptualizes rough animation in AI-based production environments not as a linear process toward completion, but as an exploratory production process in which creators’ judgments and selections are repeatedly accumulated. By shifting the focus from technical performance to internal decision-making structures and exploratory practices, this research contributes to the conceptual expansion of rough animation and provides theoretical and practical foundations for future studies on AI-driven animation tool design and production workflows.

키워드