范例推理在配色系统中的应用

The Application of CBR in color matching system

案例推理(CasebasedReasoning,CBR)是一种利用以前类似案例来理解并解决当前问题的技术。文章提出了基于CBR方法进行配色的新思想,并给出了基于CBR的印染配色系统的原型CBRMCS。文章对CBRMCS的关键技术(配方库组织结构、配方检索、配方复用、案例调整)进行了探讨,并给出了部分实验成果。选择某印染厂活性染料配方库,用CBR系统进行配色,案例检索成功率大于80%;配色后的初始配方合格率为45%,而传统方法是30%。

Case-based Reasoning (CBR) has become a successful technique for Case-based systems. This paper presents a new method, CBR,in the colour-matching area of printing and dyeing. A soft prototype system of CBRMCS is introduced. Four critical techniques in CBRMCS are discussed: Knowledge Representation,Retrieval of similar cases ,case reusing and case learning. This proposed approach is demonstrated by experiment of CBRMCS. On an database of reactive dyestuff recipes in some dyehouses, morethan 80% similar recipes can be retrieved in CBRMCS. 45% recipes are accurate in first color matching. the relevant figure of traditional colour-matching is 30%.