생성형 AI 기반한 브랜드 색채 시스템 개발 프로세스 연구

초록

This study examines the potential applications and limitations of generative AI in color scheme development. AI can participate in palette creation through two primary approaches: first, a text-driven method (Text-to-Palette) that generates multiple alternatives based on structured prompts combining brand category, target users, affective descriptors, tonal preferences, and exclusion criteria; and second, an image-driven method (Image-to-Palette) that extracts and reorganizes primary and secondary colors from reference images, posters, or natural photographs. While these approaches enhance divergent exploration efficiency, they reveal notable limitations, including insufficient cultural and contextual interpretation, weak control of systemic consistency (e.g., value/saturation ranges, contrast rules), and unstable reproducibility under identical prompts. Accordingly, this study proposes embedding designer review and secondary orientation into the Empathize, Define, and Test stages of design thinking, thereby establishing a collaborative process of “AI-driven divergence, designer-led convergence, and user validation.” This framework highlights the complementary roles of AI and human designers in brand color system development.

키워드