AIGC를 활용한 묘족 자수 이미지 디자인

초록

In the era of digital transformation, this study examines the reconstruction of visual language and design innovation pathways for Miao embroidery utilizing AIGC technology to facilitate its modern inheritance. To this end, a dataset of core Miao embroidery symbols was constructed, and generative experiments were conducted based on Stable Diffusion and LoRA models. The cultural appropriateness of the generated results was verified through a hybrid expert panel and heuristic criteria, and subsequently improved via an iterative human-computer collaboration process of "diagnosis-feedback-optimization." The findings indicate that while AIGC enhanced design efficiency, it initially showed limitations in expressing cultural context; however, retraining the models with expert knowledge significantly improved symbolic accuracy and aesthetic completeness. This study proposes a new collaborative mechanism defining AIGC as an "engine of divergent thinking" and humans as "cultural navigators," suggesting a practical methodology for the preservation and modernizationof intangible cultural heritage application.

키워드