基于眼动技术的牡丹纹样意象元素识别方法研究

Imagery Element Recognition Method of Peony Pattern Based on Eye Movement Technology

中文关键词: 传统纹样  多维尺度  眼动跟踪  均值漂移  意象识别

英文关键词: traditional pattern  multidimensional scale  eye movement data  mean shift  imagery recognition

中文摘要:

      目的 针对目前传统文化意象和现代产品设计之间匹配关系模糊、意象风格难以统一的问题，通过识别传统纹样中表征意象的图像元素，将其应用并服务于现代设计行业。方法 以宋代牡丹纹样的描述词汇为样本，采用多维尺度法归纳出意象维度，根据语义差异法完成对纹样的意象分类;基于眼动跟踪技术得到各意象维度下的注视数据，采用均值漂移算法聚类出具有意象代表性的纹样元素并量化其表征度。结果 得到高雅、祥瑞、素雅、柔润、灵动和精致6个意象维度下的纹样意象元素，并计算得出代表该意象元素在牡丹纹样意象中的贡献程度指标——表征度。结论 实验表明该方法具有可重复性和可行性，识别结果表现出了一定的意象内涵和应用价值，可为传统文化再设计提供参考思路。

英文摘要:

      There is a problem that it is difficult to match and unify the imagery and style between traditional cultural patterns and modern design. This paper aims to serve design industry by extracting elements of traditional patterns with significant imagery connotation. Firstly, based on the qualifiers of the peony patterns in the Song Dynasty, the image dimensions were summarized using multidimensional scale analysis, and the classification of patterns according to semantic differential method was completed. Secondly, the eye tracking technology was used to obtain the gazing data in each dimension, and the mean shift algorithm was applied to extract the imagery areas of interest. Finally, the quantitative representation of the pattern elements was calculated and sorted. Imagery elements in six image dimensions of elegance, exquisiteness, agility, softness, simplicity, and auspiciousness are obtained, and the representation degree, which represents the contribution degree of the imagery elements in the peony pattern imagery, is calculated. Experiments indicate that the method is repeatable and feasible. The results have certain imagery meaning and application value, which could provide reference for the re-design of traditional culture.