H.R. Giger의 조형 언어를 활용한 AI 기반 텍스타일 패턴 디자인 연구

Abstract

This study analyzes the visual language of H.R. Giger and applies it to generative AI-based textile pattern design, ultimately implementing these patterns on 3D virtual garments. Key visual elements from Giger's work—such as the biomechanical fusion of organic and mechanical forms, symmetrical structures, and metallic textures—were extracted and converted into text prompts for the image generation platform Midjourney. The AI-generated images were then post-processed in Adobe Photoshop to create seamless, tileable patterns. These patterns were applied to virtual garments, including shirts, dresses, and scarves, using CLO 3D to assess their visual effectiveness. The results indicate that the AI-generated patterns successfully captured Giger's distinctive aesthetic, demonstrating high visual precision and repeatability suitable for textile design. Variations in prompt composition resulted in different visual outcomes, and the effectiveness of the patterns varied depending on the garment type in the 3D simulations. By establishing a creative workflow that connects an artist’s unique visual language with AI-driven design, this research offers a concrete methodology with potential applications in both professional fashion practice and design education.