Dunhuang mural art provides rich visual resources for contemporary textile and apparel design, yet its application in modern products often depends on intuitive borrowing rather than structured translation. To address this issue, this study proposes a quantitative framework for translating and evaluating Dunhuang mural color and pattern features in silk scarf and evening dress design. Representative mural images were first processed using digital image analysis to extract dominant color composition, perceptual color contrast, pattern density, and orientation characteristics. These features were then aggregated into three representative visual prototypes, from which six design schemes were generated through a rule-guided translation strategy. To evaluate the design outcomes, a multi-criteria framework combining the Analytic Hierarchy Process and fuzzy comprehensive evaluation was employed. The results indicate that schemes with stronger cultural feature retention achieved higher scores in cultural expression but lower scores in practicality and perceived market acceptance because of increased visual complexity. In contrast, more balanced schemes obtained better overall performance. The proposed framework improves the reproducibility and evaluability of culturally inspired design and provides methodological support for integrating traditional visual heritage into modern textile and apparel products.

dunhuang mural, textile design, color feature extraction, pattern analysis, AHP-fuzzy evaluation