基于神经网络的覆膜产品颜色预测研究  

Research on color prediction of laminated prints based on neural network

作　　者：陈思思[1] 万晓霞[1] 孙鹏 徐宏平[1] 鞠龙[1] 

机构地区：[1]武汉大学印刷与包装系,武汉430072 [2]武汉红金龙印务股份有限公司,武汉430056

出　　处：《计算机工程与应用》2015年第4期240-243,共4页Computer Engineering and Applications

摘　　要：由于印刷品在覆膜前后存在着明显的颜色偏差,在实际生产中需要根据覆膜颜色的变化规律建立一套新的标准来指导机台生产,以保证印刷成品与客户签样的一致性。为了实现覆膜后的印刷样与客户签订的标准样之间的匹配,提出采用数码打样模拟印刷样的方式研究覆膜色彩变化规律,并利用BP神经网络建立覆膜前后色度值的映射关系,从而建立新的适用于覆膜工艺的特性文件进行数码打样,以输出的数码样来指导印刷环节,为企业提供一种低成本、高效率的色彩管理解决方案,实验数据表明该方案的可行性和准确性。Due to the obvious color deviation exists after lamination, a set of new standard is to be built to guide production according to the color change regulation before and after lamination, to make sure the color consistency between prints and approval samples. To match the laminated prints and approval samples, the color change rule of lamination is researched with digital proofing. The mapping relation of chromatic values before and after lamination is set up by BP neural network.Meanwhile, the new ICC profile is built for digital proofing, and the proofing sheets an guide printing process. It may provide relative printing enterprises with a low-cost and high-efficiency color management solution and the statistic shows its feasibility and accuracy.

关 键 词：覆膜 颜色预测 颜色偏差 神经网络 数码打样 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]