A Multimodal Incentive-Based Approach to Anime Character Image Design

A Multimodal Incentive-Based Approach to Anime Character Image Design

연구 분야: Databases

학회: 2024 5th International Conference on Intelligent Design (ICID)

The rapid development of multimodal large language models has transformed knowledge-based incentives into a key source of inspiration for designers. As traditional design workflows centered on personal creativity and online resources shift toward data-driven innovation, it is crucial to explore how multimodal data can enhance creativity and efficiency. This paper introduces a multimodal incentive-based approach for anime character design, supported by the construction of a Chinese anime character database and a design process integrating both textual and visual data. Experimental results demonstrate that multimodal incentives significantly improve creative stimulation, design efficiency, and innovation for designers.