没有合适的资源?快使用搜索试试~ 我知道了~
Spectral-based color separation method for a multi-ink printer
0 下载量 51 浏览量
2021-02-25
23:11:45
上传
评论
收藏 419KB PDF 举报
温馨提示
A spectral-based 8-ink characterization model is developed to accurately predict the recipe for a multi-ink printer. The 8-ink color separation method is a union of five 3-ink and six 4-ink combinations based on the cellular Yule-Nielsen spectral Neugebauer model with a recipe selection strategy. The performance levels of the forward and backward models are evaluated for individual ink combinations using printed testing samples. Furthermore, the spectral-based method performs better compared wit
资源推荐
资源详情
资源评论




















COL 9(6), 063301(2011) CHINESE OPTICS LETTERS June 10, 2011
Spectral-based color separation method for
a multi-ink printer
Binyu Wang (
QQQ
)
1,2
, Haisong Xu (
MMM
°°°
ttt
)
1
∗
, M. Ronnier Luo
2
, and Jinyi Guo (
HHH
AAA
)
1,2
1
State Key Laboratory of Modern Optical Instrumentation, Zhejiang University, Hangzhou 310027, China
2
Department of Colour Science, University of Leeds, Leeds, LS2 9JT, UK
∗
Corresponding author: chsxu@zju.edu.cn
Received November 12, 2010; accepted December 17, 2010; posted online April 28, 2011
A sp ectral-based 8-ink characterization model is developed to accurately predict the recipe for a multi-ink
printer. The 8-ink color separation method is a union of five 3-ink and six 4-ink combinations based on
the cellular Yule-Nielsen spectral Neugebauer model with a recipe selection strategy. The performance
levels of the forward and backward models are evaluated for individual ink combinations using printed
testing samples. Furthermore, the spectral-based method performs bett er compared with the XYZ-based
approach. On the b asis of the backward model performance, a novel fast recipe selection strategy is
proposed and estimated.
OCIS codes: 330.1690, 300.6550, 330.1710, 330.1715.
doi: 10.3788/COL201109.063301.
In printer systems, two major approaches are used, i.e.,
the analytical and generic methods, to derive a character-
ization model. The analytical metho d is based on mod-
eling physical phenomena, such as Kubelka-Munk the-
ory and Neugebauer equations. The parameters of the
model are usually obtained from a relatively small num-
ber of samples, while the accuracy depends on the e x-
tent to which the model describes the actual behavior of
the device; some of these models are difficult to invert.
The generic method requires a large set o f training data
to build the model, as in three-dimensional (3D) look-
up table with interpolation
[1]
. Because the spectral data
contain all color information compared with tristimulus
values, the spectral-based approach achieves higher ac-
curacy and lower illuminant metamerism than does the
XYZ-based method at the expense of more calculations.
Typical 4-ink color printers have sma ll color gamut and
limited deg rees of freedom; thus, the use o f more inks not
only incre ases color gamut and the number of degrees of
freedom, but also improves accuracy
[2,3]
.
In this letter, the cellular Yule-Nielsen spectral Neuge-
bauer (CYNSN)
[2,4]
model is implemented to character-
ize a HP Desig njet Z3200ps 44 -inch photo printer, which
uses 12 inks
[5]
. For s implicity, 8 inks are used in the
exp eriment, including grey (Gy), photo black (K), light
cyan (C), magenta (M), yellow (Y), chromatic red (R),
green (Gn), and blue (B). The 8-ink color separation al-
gorithm is a union of five 3-ink and six 4-ink sets based
on the CYNSN model
[6]
with a recipe selection strat-
egy, containing the ink combinations of CMY, RMY,
CGnY, CMB, RGnB, CMYK, RMYK, CGnYK, CMBK,
RGnBK, a nd CMYGy. A large number of training and
testing hardcopy samples were prepared and measured,
and then the performa nce levels of the forward and back-
ward models were individually evaluated. In addition,
the spec tral-based method was compared with the XYZ-
based approach in terms of color difference, reflectance
root mean square
[7]
(RRMS) error, and metamerism in-
dex (MI). Finally, a novel fast recipe selection strategy
with high accuracy is proposed for the printer.
In 1937, Neugebauer developed a method for modeling
CMY halftone printing
[4]
. It was based on improving the
Murray-Davies
[2]
model and can be extended to use mor e
inks. Afterward, two main improvements
[2]
were pre -
sented. One is the Yule-Nielsen n-factor describing the
light sca ttering. The classical CMY Neugebauer model
modified by the Yule-Nielsen effect is expressed as
R(λ) =
(
8
X
i=1
w
i
[R
i
(λ)]
1/n
)
n
, (1)
where R
i
(λ) is the reflectance of the ith primaries (i =
1, 2, · · ·, 8) at wavelength λ, w
i
denotes the weighting
coefficient for the ith primary defined by the Demichel
equation
[8]
, n-factor is usually determined by optimiza-
tion, and the typical value of n-factor rang e s from 1 to
4. The other improvement is called the cellular Neuge-
bauer model. Firstly, the entire color space (e.g. CMY,
CMYK) is divided into smaller sub-spaces called ce lls,
and then the classical Neugeba uer model is applied in
the selec ted cell dependent on the input ink amounts.
Combining these two improvements with the classical
Neugebauer model forms the CYNSN model, in which
predicting ac c uracy drama tically increases.
The Yule-Nielsen spectral Neugebauer (YNSN) model
is not analytically invertible; thus, its inversion is usually
solved using iteratio n. A well-known method was pr o-
posed by Urban and Grigat, in which linear regres sion
iteration was used
[9]
, based on which Li et al. introduced
“QR” decomposition, further reducing computational
costs
[10]
. Therefore, Li algorithm was employed in this
letter.
Before conducting the backward YNSN model, deter-
mining the optimal cell is an indispensable step, within
which the most accurate recipe can be predicted by the
backward model. Urban et al. developed a cellular lin-
ear regression iteration algorithm, which dramatically
decreased the number of cells tr ied but did not guaran-
tee that the selected cell was the actual optimal cell
[11]
.
Later, Guo et al. proposed a cell searching method,
which ensured that the optimal cell was found at the
1671-7694/2011/063301(4) 063301-1
c
2011 Chinese Optics Letters
资源评论


weixin_38621870
- 粉丝: 7
- 资源: 936
上传资源 快速赚钱
我的内容管理 展开
我的资源 快来上传第一个资源
我的收益
登录查看自己的收益我的积分 登录查看自己的积分
我的C币 登录后查看C币余额
我的收藏
我的下载
下载帮助


最新资源
- 人工智能&深度学习:PyTorch 实战 - DataLoader 自定义花卉数据集制作资源(源码+数据集+说明)
- 研发效能提升全解析:业界顶尖专家答疑效能度量与实践
- 素材图片文档合集-新空间
- 安徽省创新型中小企业名单(2024年度第二批)
- DeepSeek完全实用手册V1.0+-+至顶AI实验室.pdf
- ihrm接口测试postman脚本
- 汇编语言教程&案例&相关项目资源
- Comsol仿真解析纳米孔超表面的手性响应及其应用探究,纳米孔超表面的手性响应与COMSOL仿真的探讨分析,comsol仿真纳米孔超表面的手性响应 ,关键词:comsol仿真; 纳米孔超表面; 手性响
- 哪吒之魔童闹海-ts文件
- 【火绒-2025研报】“火绒终端安全管理系统1.0版”.pdf
- 【Workday-2025研报】探索银行中AI驱动技术转型的投资回报率。.pdf
- 【科智咨询-2025研报】deepseek对算力产业的影响.pdf
- 【未知机构-2025研报】3D 打印行业发展研究报告.pdf
- 【莱坊-2025研究报告】Melbourne CBD Office Market February 25.pdf
- 【YOYI悠易-2025研报】突破传统奢侈品营销模式:数变奢侈品营销新增长.pdf
- 【360-2025研报】2025年1月勒索软件流行态势分析.pdf
资源上传下载、课程学习等过程中有任何疑问或建议,欢迎提出宝贵意见哦~我们会及时处理!
点击此处反馈



安全验证
文档复制为VIP权益,开通VIP直接复制
