AI-Mediated Workflow Development for Controllable Textile Design Generation
- 문서 유형
- 연구출판물 · 학술지 논문
- 등재
- KCIScopus
- 저널 순위
- Scopus Q2KCI Q2
- 저널
- 한국의류학회지50(2) · pp.440 · 2026
- 피인용
- Scopus 0KCI 0
- FWCI (Scopus)
- 0
- ISSN
- 1225-1151
저자
초록
This study examines whether the variability in outcomes from generative AI–based textile design can be transformed into a manageable design process through structured prompt design. To examine the validity of the process, experiments were conducted using a specific case. Unstructured planning-stage textual inputs were transformed into structured design language units. Three staged experiments were conducted under identical motif inputs and generation conditions, with increasing prompt structural explicitness. Generated images were evaluated using predefined structural usability criteria that distinguish mechanical repeat continuity from functional repeat validity. A clear reduction in the discrepancy was observed between seam-level tiling success and functional repeat applicability as structural parameters were integrated into the prompt. Free-form descriptive prompts often satisfied mechanical continuity but frequently failed to meet the minimum distributional conditions required for textile usability. Conversely, parameter-explicit prompt configurations demonstrated systematically higher rates of structural validity rates under repeated generation. After separating fixed baseline inputs from adjustable structural parameters, a large-scale modular validation proved repetition stability. These findings clarify how structured prompt configuration serves as a controllable design interface that stabilizes structural variables within iterative AI-mediated textile workflows.
키워드
Generative AIPrompt structuringStructural prompt configurationTextile designWorkflow
식별자
- Scopus EID
- 2-s2.0-105044671670
- KCI
- ART003334581



로그인
연구물 목록으로