Characteristic Reverse Algorithm Based on Multi-Illuminants Printer

Yan Zhao; Yang Xu; Cheng Gao; Changjun Li

doi:10.3788/LOP57.153302

Laser & Optoelectronics Progress, Volume. 57, Issue 15, 153302(2020)

Characteristic Reverse Algorithm Based on Multi-Illuminants Printer

Yan Zhao, Yang Xu^*, Cheng Gao, and Changjun Li

Author Affiliations

School of Computer and Software Engineering, University of Science and Technology, Anshan, Liaoning 114051, China

show less

Abstract Get PDF(in Chinese)

Figures & Tables(6)

Fig. 1. Three-level Cellular Neugebauer model

Download full size

Table 1. Eight primary colors, ink combinations, reflectance, and Demchel equations for the Neugebauer model

View table

Table 1. Eight primary colors, ink combinations, reflectance, and Demchel equations for the Neugebauer model

Primary	c,m,y	Reflectance	Demchel equation
White	(0,0,0)	R₁(λ)	a₁=(1-c)(1-m)(1-y)
Cyan	(1,0,0)	R₂(λ)	a₂=c(1-m)(1-y)
Magenta	(0,1,0)	R₃(λ)	a₃=(1-c)m(1-y)
Yellow	(0,0,1)	R₄(λ)	a₄=(1-c)(1-m)y
Red	(0,1,1)	R₅(λ)	a₅=(1-c)my
Green	(1,0,1)	R₆(λ)	a₆=c(1-m)y
Blue	(1,1,0)	R₇(λ)	a₇=cm(1-y)
Black	(1,1,1)	R₈(λ)	a₈=cmy

Table 2. Comparison of five algorithms iterative multiplications

View table

Table 2. Comparison of five algorithms iterative multiplications

Method	4-ink iterative multiplications per iteration
UG	M₁ = 4 [8(vector evaluation)+ 16×L(matrix×vector)+ 2×L(inner product)+1]
Urban-SVD	${M_{2}}_{}$ = 4 [8(vector evaluation)+ 16×16(matrix×vector)+ 2×16(inner product)+1]
LL-QR	M₃ = 4 [8(vector evaluation)+ (16×17)/2(matrix×vector)+ 2×16(inner product)+1]
TSV-based	M₄ = 4 [8(vector evaluation)+ 16×3(matrix×vector)+ 2×3(inner product)+1]
ML-mix	M₅ = 4 [8(vector evaluation)+ 16×6(matrix×vector)+ 2×6(inner product)+1]

Table 3. Forward spectral prediction model color difference and spectral root mean square error
View table
Table 3. Forward spectral prediction model color difference and spectral root mean square error
Model Paper Level ΔE_D50,Ave ΔE_D50,Max ΔE_D50,Med ΔR_Ave
cmy Unknown 5 1.4097 8.0832 0.8148 0.0551
9 0.9614 6.2115 0.6154 0.0338
cmyk art 5 0.7333 2.4563 0.6609 0.0260
hm 5 1.0010 6.3253 0.7445 0.0217

Table 4. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level and 9-level CYNSN model

View table

Table 4. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level and 9-level CYNSN model

Level	Method	ΔE_Ave				ΔR_Ave	M_MIA,D50	M_MIF11,D50	M_MID65,D50
Level	Method	D50	D65	A	F11	ΔR_Ave	M_MIA,D50	M_MIF11,D50	M_MID65,D50
	LL-QR	0.5130	0.5328	0.4622	0.5419	0.0228	0.1314	0.1754	0.0645
	D50	0.3271	0.3461	0.3562	0.4083	0.0250	0.1304	0.1733	0.0650
5	D50D65	0.3301	0.3430	0.3716	0.4205	0.0254	0.1296	0.1748	0.0648
	D50A	0.3695	0.3931	0.3405	0.4042	0.0242	0.1285	0.1725	0.0616
	D65A	0.3675	0.3867	0.3553	0.4131	0.0248	0.1291	0.1744	0.0616
	D50D65A	0.3438	0.3630	0.3690	0.4282	0.0251	0.1292	0.1747	0.0632
	LL-QR	0.2952	0.314	0.2562	0.2903	0.0151	0.0886	0.1194	0.0435
	D50	0.1676	0.1919	0.1975	0.2345	0.0171	0.1006	0.1405	0.0500
9	D50D65	0.1626	0.1800	0.2096	0.2394	0.0172	0.1035	0.1441	0.0517
	D50A	0.2046	0.2267	0.1827	0.2150	0.0165	0.1027	0.1355	0.0480
	D65A	0.1972	0.2162	0.1913	0.2172	0.0166	0.1017	0.1368	0.0473
	D50D65A	0.1713	0.1908	0.2003	0.2330	0.0170	0.1018	0.1377	0.0488

Table 5. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level CYNSN model with two papers

View table

Table 5. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level CYNSN model with two papers

Level	Paper	Method	ΔE_Ave				ΔR_Ave	M_MIA,D50	M_MIF11,D50	M_MID65,D50
Level	Paper	Method	D50	D65	A	F11	ΔR_Ave	M_MIA,D50	M_MIF11,D50	M_MID65,D50
5		LL-QR	0.7378	0.7643	0.6867	0.9073	0.0252	0.2140	0.4105	0.0885
		D50	0.6501	0.7047	0.8554	1.0893	0.0384	0.3843	0.6200	0.1474
	art	D50D65	0.6644	0.6915	0.8569	1.0820	0.0362	0.3493	0.5792	0.1338
		D50A	0.6475	0.7021	0.6378	0.8540	0.0310	0.2869	0.4993	0.1112
		D65A	0.6497	0.6935	0.6801	0.8847	0.0307	0.2741	0.4782	0.1079
		D50D65A	0.6123	0.6369	0.7055	0.9083	0.0294	0.2469	0.4606	0.1006
		LL-QR	0.8883	0.8893	0.8739	1.0263	0.0225	0.2307	0.5228	0.0942
		D50	0.9142	0.9475	0.9864	1.1743	0.0325	0.4068	0.7409	0.1458
	hm	D50D65	0.9011	0.9178	0.9880	1.1653	0.0316	0.3926	0.7241	0.1391
		D50A	0.8417	0.8869	0.7746	0.9783	0.0271	0.3360	0.6772	0.1195
		D65A	0.8777	0.9044	0.8379	1.0274	0.0264	0.3203	0.6528	0.1132
		D50D65A	0.8444	0.8685	0.8703	1.0631	0.0269	0.3235	0.6610	0.1156

Tools

Get Citation

Copy Citation Text

Yan Zhao, Yang Xu, Cheng Gao, Changjun Li. Characteristic Reverse Algorithm Based on Multi-Illuminants Printer[J]. Laser & Optoelectronics Progress, 2020, 57(15): 153302

Download Citation

EndNote(RIS)BibTex Plain Text

Set citation alerts for article

Save article for my favorites

Paper Information

Category: Vision, Color, and Visual Optics

Received: Oct. 25, 2019

Accepted: Nov. 29, 2019

Published Online: Aug. 4, 2020

The Author Email: Yang Xu (xuyang_1981@aliyun.com)

DOI:10.3788/LOP57.153302

Topics

laser devices and laser physics

Lasers and Laser Optics

Laser physics

laser manufacturing

Instrumentation, Measurement and Metrology

Table 1. Eight primary colors, ink combinations, reflectance, and Demchel equations for the Neugebauer model

Table 1. Eight primary colors, ink combinations, reflectance, and Demchel equations for the Neugebauer model

Table 2. Comparison of five algorithms iterative multiplications

Table 2. Comparison of five algorithms iterative multiplications

Table 3. Forward spectral prediction model color difference and spectral root mean square error

Table 3. Forward spectral prediction model color difference and spectral root mean square error

Table 4. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level and 9-level CYNSN model

Table 4. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level and 9-level CYNSN model

Table 5. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level CYNSN model with two papers

Table 5. When n=3, comparison results of color difference, spectral root mean square error, and metameric index for each source of 5-level CYNSN model with two papers