现代电子技术2018，Vol.41Issue(5)：74-77,4.DOI:10.16652/j.issn.1004-373x.2018.05.017
参考图像集的鲁棒性层次结构模型
Robust hierarchical structure model for reference image sets
摘要
Abstract
Aiming at the multi-factor qualitative decision problem for color reference image sets establishment in color resto-ration of cultural relics,a color restoration hierarchical structure model based on analytic hierarchy process(AHP)is proposed. The historical data records,expert experience and spot investigation are combined to analyze the actual color fading problem, summarize the ontology elements representing the color of cultural relics,and study the intrinsic connection between ontology elements and fading. The expert group decision is used to establish the color restoration hierarchy model following the steps of expert group decision-making,color elements collection,knowledge discovery,main influence factor determination,hierarchical structure analysis,expert judgment matrix establishment,judgment matrix solving and color restoration hierarchical structure model establishment based on weight. The two-level filtering method is used to solve the model layer-by-layer and judge the matrix weight to realize the reference image sets'search system based on the model.The experimental results show that the method can accurately find the reference image combination for the faded images,and provide the scientific reference basis for the subsequent color restoration.关键词
文物数字化保护/层次结构模型/群组决策/判断矩阵/色彩复原/颜色迁移Key words
digital preservation of cultural relic/hierarchical structure model/group decision-making/judgment matrix/color restoration/color transfer分类
信息技术与安全科学引用本文复制引用
李娜,覃锐华,王特,张栋,宋悦..参考图像集的鲁棒性层次结构模型[J].现代电子技术,2018,41(5):74-77,4.基金项目
国家级大学生创新创业训练计划项目(201610704022) （201610704022）
西安科技大学培育基金(201744) （201744）
西安科技大学博士启动金(6310116057)Project Supported by College Students′ Innovation and Entrepreneurship Training Program(201610704022),The Research and Development Fund of Xi'an University of Science and Technology(201744),The Doctoral Fund of Xi'an University of Science and Technology(6310116057) （6310116057）
