摘要
提出了一种用于彩色打印机色空间变换的方法。基于输出色彩单色色度特性与数字驱动值以及复色与单色色度特性之间存在着一维和三维的特征关系,通过色度密度的概念及神经网络方法将此特征关系定量化,由此建立了输出色度色空间与输入数字色空间之间的变换关系。采用729组训练数据的实验结果表明,该色度密度方法达到的变换精度高于一些直接建立输出输入关系的方法。喷墨打印系统的实验结果表明,输出CIELAB色度平均和最大色差分别为3.18和16.0;260组泛化测试CIELAB色度平均和最大色差分别为4.00和16.8。
A new method is proposed for color space conversion of color printers. Based on the concept of colorimetric density and artificial neural network, the one-dimension relationship between colorimetric characteristics of primary colors and printing data and the three-dimension colorimetric relationship between complex colors and primary colors were established together with the relationship between the two color spaces, output colorimetric space and input data space. Experiments were made with 729 samples and the experimental results showed that the conversion accuracy of this method is higher than those of some other methods, which build direct relationship between output and input. For one printer, the mean and maximal CIELAB c.d. (color difference) of 729 samples are 3.18 and 16.0, respectively, and these of 260 test samples are 4.00 and 16.8.
出处
《光学精密工程》
EI
CAS
CSCD
2004年第3期265-269,共5页
Optics and Precision Engineering
关键词
彩色打印机
色空间变换
色度密度
多项式拟合
人工神经网络
色差
colour printer
color space conversion
colorimetric density
polynomial fit
artificial neural network
color difference
