Scholarlink

<p>Color is a key visual element of heritage brocade, but traditional extraction relies on manual sampling and subjective judgment. This study proposes a computational framework for heritage brocade color extraction and analysis to support systematic and sustainable preservation. The model integrates rolling guidance filtering for texture suppression, adaptive clustering for single-image color extraction, hierarchical agglomerative clustering with CIEDE2000 for global color reduction, Natural Color System representation, and Apriori-based association rule mining for color co-occurrence analysis. Taking Chinese Tujia brocade as a case study, 285 high-quality images were analyzed. Comparative experiments showed that Mean Shift achieved the best overall color extraction performance with acceptable computational cost. The analysis yielded 68 representative characteristic colors and revealed meaningful co-occurrence patterns. Expert evaluation and recoloring applications confirmed that the results aligned with traditional aesthetics and visual perception. The framework provides data-driven support for systematic documentation, analysis, and sustainable utilization of heritage brocade colors.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Color extraction and analysis of heritage brocade through computational algorithms

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月9日

查看全文
  • Xiaoyan Li,
  • Tao Ma,
  • Hong Zhao,
  • Yu Liu

摘要

Color is a key visual element of heritage brocade, but traditional extraction relies on manual sampling and subjective judgment. This study proposes a computational framework for heritage brocade color extraction and analysis to support systematic and sustainable preservation. The model integrates rolling guidance filtering for texture suppression, adaptive clustering for single-image color extraction, hierarchical agglomerative clustering with CIEDE2000 for global color reduction, Natural Color System representation, and Apriori-based association rule mining for color co-occurrence analysis. Taking Chinese Tujia brocade as a case study, 285 high-quality images were analyzed. Comparative experiments showed that Mean Shift achieved the best overall color extraction performance with acceptable computational cost. The analysis yielded 68 representative characteristic colors and revealed meaningful co-occurrence patterns. Expert evaluation and recoloring applications confirmed that the results aligned with traditional aesthetics and visual perception. The framework provides data-driven support for systematic documentation, analysis, and sustainable utilization of heritage brocade colors.