748 CHINESE OPTICS LETTERS / Vol. 7, No. 8 / August 10, 2009
Color reproduction from desktop display to projector based
on visual matching
Weijun Han (
¸¸¸
¥¥¥
), Junsheng Shi (
ddd
)))
)
∗
, Ping He (
ååå
±±±
), and Lijun Yun (
|||
)
Color & Image Vision Lab, Yunnan Normal University, Kunming 650092, China
∗
E-mail: shi-js@263.net
Received December 5, 2008
We present a novel method of color reproduction from desktop displays to projectors via visual assessment.
The model is based on visual matching nine color patches between a display and a projector. The effects
of the method to improve color reprodu ction are tested for 30 samples by visual and color difference
evaluations. The expeirmental results of visual evaluation show that the color reproduction is improved
by 87.5%. The maximum, minimum, and average color d ifferences between the displayed colors and
the projected ones before and after correction are 28.94, 4.35, 16.78, 16.51, 0.64, and 3.51 ∆E
∗
ab
units
respectively, which are consistent with the results of visual evaluation.
OCIS codes: 330.1720, 330.1730.
doi: 10.3788/COL20090708.0748.
It has become popular that color images and graphs
are displayed on computer screens , and furthermore pro-
jected on the large white screen or wall by projectors.
Color reproduction from displays to projectors is im-
portant in so me cases. However, it often happens that
the projected colors by a projector look different from
the ones on a display screen. Generally, color man-
agement system (CMS) for achieving cross-media repro-
duction needs color measurement instruments for de-
vice characterizations
[1,2]
, color appearance models, and
gamut mapping
[3,4]
. However, this is for professionals
(such as graphic arts industry) and color critical use rs
(such as medical imaging), and the color measure ment
instruments are expensive and also no t conve nient for
general users.
With the publication of the color appearance model
CIECAM02 in 2004 by the International Commissio n on
Illumination (CIE)
[4]
, the model for cro ss-media color
reproduction is becoming more and more accurate and
complex. A great deal of researches on cross-media color
reproduction have been concentrated on the case between
computer displays and printers. There are much less re-
searches on color reproduction between displays and pro-
jector. The studies on the projected color of pr ojectors
were focused on color characterization, display character-
istic, optimization of tiled display color uniformity, and
color gamut
[5−10]
.
In this letter, we present a sample method to match
several dis palyed and projected color patches by human
vision without using any color measurement instrument.
The transformation between two devices in color space
is then attained by implementing simple mathematical
optimization based on matching color a ppea rance in the
CIELAB color s pace. The method is tested by visual and
color difference evaluations.
One color with RGB (red, green, and blue) values in
the computer video memory is displayed on a display
screen and simultaneously projected on the large white
screen/wall by a projector with the same RGB values. In
general cases, there is the difference in color appearance
between displayed color and projected one, as s hown in
Fig. 1, which is the case of no matching color appearance
in the C IELAB color space.
To achieve color matching between pr ojected and dis-
played one, the RGB values of the color on the display
screen should be corrected or converted to R
′
G
′
B
′
values
projected by the projector, as shown in Fig. 2.
The principle and method to attain R
′
G
′
B
′
from RGB
are described as follows.
A simple model of display characterization is shown
as
[1,2]
"
X
Y
Z
#
=
"
a
11
a
12
a
13
a
21
a
22
a
23
a
31
a
32
a
33
#"
(R/25 5)
γ
dis
(G/255)
γ
dis
(B/255)
γ
dis
#
= A ×
"
(R/25 5)
γ
dis
(G/255)
γ
dis
(B/255)
γ
dis
#
, (1)
where γ
dis
is the dis play gamma, and A is the linear ma-
trix of co lorimetric characterization for the display. At
the same time, according to the principle of additive color
for a projector, a simple model of projector characteriza-
Fig. 1. The case of no matching between projected color and
displayed color.
Fig. 2. The case of matching between projected color and
displayed color.
1671-7694/2009/080748-05
c
2009 Chinese Optics Letters