Optimizing the conceptual design process of animation scenes using generative AI

초록

Generative AI, with its rapidly developing technologies, has changed how designers look to solve problems relating to the scene's animation: eliminating inefficiencies, reducing high production costs, and lifting the cognitive load involved in the conception of an animation scene. Therefore, a systematic conceptual design model based on the design of animation scenes, which integrates generative AI tools-ChatGPT and Midjourney-is proposed in this paper. Based on the Co-evolution Model of Design and Divergent-Convergent Thinking, this model enables iterative co-evolution between the problem space and solution space, further provisionally refining a designer's capacity to explicate goals while exploring various visual solutions effectively. The following is a series of design iterations in which such a model has been applied practically in line with the Research through Design approach. The results thus demonstrated that the proposed framework reduces the cognitive load on designers, increases efficiency, and improves creativity and coherence in the scenes created. Hence, this contributes to AIGen-assisted design-valid methodologies from both theoretical and practical perspectives, appealing to both practitioners and academics with the quick turn of generative AI into creative practice.

키워드