Adaptive color reproduction method to various user s monitor environment in color printer

Transcription

1 Adaptive color reproduction method to various user s monitor environment in color printer Dae-Chul Kim a, In-Su Jang a, Chang-Hwan Son b, Kyung-Man Kim b and Yeong-Ho Ha a a School of Electrical Engineering and Computer Science, Kyungpook National Univ., 1370, Sankyuk-dong, Buk-gu, Taegu , South Korea; b Digital Printing Division, Samsung Electronics Co. Ltd. 416 Maetan-3 Dong, Yeongtong-Gu, Suwon-City, Gyeonggi-Do, South Korea ABSTRACT The Color Management System in recent printers adopts ICC profiles for both monitors and printers. However, the ICC profile doesn t contain the characteristics of reproduced color on each monitor, because the color on each monitor is changed by user adjustment such as color temperature, brightness, and contrast adjustment. It is also depended on the backlight type and lifetime. As a result, unwanted color is reproduced on the printed paper, not like that on the monitor. To overcome the color difference between monitors and printers, we need to control the information of ICC profile. That is, first, the ICC profile is generated by the measurement of monitors having user set, then, through the CMS, the color on monitors can be produced on printed paper. However, it is difficult to apply the above system for normal users due to absence of measuring equipment and time consuming process. Therefore, this paper proposes a novel color matching technique based on the estimation of condition for each monitor having user set. The estimation is performed by a simple comparison visual test using a test image on printed paper and monitor. Then, the condition of monitor is applied to the ICC profile. As a result, the new ICC profile contains the color difference between user monitor and printer. The experimental results show the printed images using our proposed method have almost similar color with those on monitors. Keywords: Color Management System, ICC profiles, Monitor, Printer, Color Temperature, Color matching, Contrast, User s environment 1. INTRODUCTION In general, the device characterization and gamut mapping processes are necessary to match the colors between two color imaging devices. Because the color spaces used in each device are different, the color data should be converted into a device-independent color space, such as CIEXYZ or CIELAB. 1,2 Also, gamut mapping should be applied because of the difference of color gamut. There are many device characterization methods and gamut mapping methods. 3,4 Thus the International Color Consortium defined the profile format for this color matching process. The ICC profile provides color management system with the information necessary to convert color data between native device color spaces and device-independent color space. Thus the color management system can easily convert color data with this ICC profile. 5 Especially, in printing system, ICC profile is very useful because the color space of printer is CMY color space contrary, as opposed to RGB color space of monitor, and the difference of gamut with monitor is bigger comparing other output devices. However, the color rendition of a monitor is not always fixed. The backlight color of an LCD monitor will change over time, so its ICC profile is not always consistent with the color characteristic of the device. Moreover, usually the use can control the monitor color temperature, brightness and contrast values. After the user has modified such settings, the monitor ICC profile is no longer valid. In contrast with monitor, color rendition of a printer is less prone to change because the toner or ink are replaceable with new ones. Yet, once we are able to produce an ICC profile that matches the user s monitor characteristics by measuring them, then the color management system becomes able to yha@ee.knu.ac.kr; phone ; fax ; Color Imaging XV: Displaying, Processing, Hardcopy, and Applications, edited by Reiner Eschbach, Gabriel G. Marcu, Shoji Tominaga, Alessandro Rizzi, Proc. of SPIE-IS&T Electronic Imaging, SPIE Vol. 7528, 75280Q 2010 SPIE-IS&T CCC code: X/10/$18 doi: / SPIE-IS&T/ Vol Q-1

2 correctly perform color matching. Unfortunately this method is of difficult application, because in general the measuring equipment is not available and, even then, it takes a long time and new measurements according to monitor color temperature, brightness and contrast. Thus, calibration of the monitor is generally performed before the color reproduction process between monitor and printer. After color of monitor is controlled comparing the ICC profile or standard color space, the color matching is performed. It can get comparatively accuracy results. But measuring instruments are expensive and the calibration process is inconvenient. Monitor calibration without a measuring instrument induces bigger color difference between monitor and printer. 6,7 Usually a user corrects the color of an image with graphic tools while watching the monitor, and wants to get a printed copy that looks the same. In order to grant the best possible color match, color features of the monitor should be estimated at printing time. Thus, in this paper we propose a color matching method using a simple visual test. We first start by modeling the monitor characteristics according to a variety of monitor s color temperature, brightness and contrast. The input-output luminosity curve and RGB primary chromaticity in relation to the variation in brightness, color temperature and contrast can then be estimated. In the end, we produce a corrected ICC profile and apply it to the color management system. 2. COLOR CHARACTERISTICS OF MONITORS 2.1 User s monitor environment Usually the color space for newly produced monitors is set to srgb. The chromaticity values of the RGB primaries are similar to srgb ones and the gamma of tone curves is closed to 2.2. White point is also determined by white chromaticity of standard illuminant, D65. Because of differences in material and production methods the color rendition varies between monitor and monitor. The color space might be changed according to the users taste or vendor intent. Moreover, the color rendition of a monitor is changed over time because the most common LCD backlighting is given by fluorescent light tubes. As shown in Figure 1, even if two monitors share the same model number, their different age leads to different color rendition. Even in the case in which the date of manufacture is also the same, monitors reproduce the different colors if the user s monitor settings, such as color temperature, contrast, and brightness, is different. (a) (b) Figure 1. Test image displayed on two samples of the same monitor. 2.2 Color temperature A monitor s color temperature is generally set as a 6500K, the same as the white chromaticity of srgb color space. However, the white point of most monitors has a different chromaticity than that of a 6500K white. Figure 2 shows the measured white chromaticity of the monitors used in the tests, measured by a spectrophotometer. The white points are distributed around the locus of standard illuminants. In the case of an old monitor, for example, the color temperature is closed to 5000K. Other monitors have a white point close to 7000K. SPIE-IS&T/ Vol Q-2

3 Figure 2. White point chromaticity of various monitors plotted in CIE-xy. Furthermore, there can be additional color temperature modes, such as 9300K or 5000K, which some monitors represent as cool and warm modes. Figure 3 shows how an image is affected by changes of the white point temperature. For higher color temperatures, the image is shifted towards blue. For lower color temperatures, the image is shifted towards yellow. Figure 3. Color shift of monitor varying color temperature. y y x Figure 4. Chromaticities of red, green, blue primaries and white point for several monitors, plotted in CIE-xy. x SPIE-IS&T/ Vol Q-3

4 Table 1. CIExy values of RGB and white channel. Channel srgb Average St. dev. X Y X y X Y Red Green Blue White Difference of represented colors for each monitor does not mean the difference of chromaticity of RGB primary. In other words, even if the color rendition of two monitors differs, they still might have the same RGB primaries. color filter and back light unit of a monitor is not changed for varying the color temperature mode. As shown in the Figure 4, the chromaticity values of the red, green and blue channels are clustered around the RGB points in the srgb color space. Instead, white points shows a higher spread, even if the average color of several monitors is similar to the white of srgb color space. Instead of the chromaticity values for RGB channels, the absolute luminance values of RGB channel is different. The color of monitor for varying the color temperature mode is changed by the luminance of RGB channel as shown in Table 2. As a result, the variation of luminance for RGB channel can be estimated by RGB gain as follows; R G B R G B Xe Xe X e XsRGB XsRGB X srgb gr 0 0 R G B R G B Ye Ye Ye YsRGB YsRGB Y srgb 0 gg 0 = R G B R G B Z 0 0 g e Ze Ze ZsRGB ZsRGB Z srgb B (1) Figure 5. White chromaticity shift of monitor by varying color temperature RGB gains are multiplied by the XYZ values of the srgb primaries, so the intensity of R, G and B is increased linearly. The RGB gains are applied independently for each channel. To estimate RGB gains, the measured XYZ values for RGB channel of monitor are used as follows R G B 1 R G B srgb srgb srgb monitor monitor monitor gr 0 0 X X X X X X R G B R G B 0 gg 0 = YsRGB YsRGB YsRGB Ymonitor Ymonitor Ymonitor. (2) 0 0 g R G B R G B B ZsRGB ZsRGB ZsRGB Zmonitor Zmonitor Zmonitor SPIE-IS&T/ Vol Q-4

5 Table 2. Change of luminance values for RGB channel by color temperature mode Model Monitor 1 Monitor 2 Monitor 3 Monitor 4 Monitor 5 Channel Luminance Default Cool or 9300K Variation Red Green Blue Red Green Blue Red Green Blue Red Green Blue Red Green Blue Contrast and brightness The relationship between digital input values and output luminance levels for the red, green, and blue channels is not linear. Thus, tone curve models, such as GOG and S-curve, are generally used to estimate such relationship. We found the GOG model to be a better approximation of the tone curves of the test monitors. The GOG model is represented by the following equation. d Y = k + k 2 1 i i g, i N o, i r i (3) The gamma parameter has a higher influence on the result than gain and offset. However, when varying the contrast mode of the monitor, as shown in Figure 6(a), the estimated gamma values do not change. Instead, the estimated gain values increase with the contrast value. According to this phenomenon, the variation of contrast can be estimated as a gain value. On the contrary, the tone curves do no change when there is a variation in brightness: only the absolute luminance value changes. Furthermore, the information stored in ICC profiles is relative to a normalized maximum. Therefore we did not consider brightness in the color matching process. In addition, variations produced in the tone curves by modifications in brightness and contrast are orthogonal. As a consequence, the GOG model is used as the tone curve model of the monitor, and the change of the tone curve is modeled through the estimation of the gain value for contrast. Notably, the information regarding the tone curve is stored as normalized integer values in the ICC profile. Thus, the information for the absolute luminance is not stored, and gamut mapping and characterization process does not need it. SPIE-IS&T/ Vol Q-5

6 (a) (b) Figure 6. Monitor tone curves, (a) tone curves for contrast modes, (b) tone curves for brightness modes. 3. SOFT PROOFING FOR ESTIMATION OF MONITOR CHARACTERISTIC 3.1 Soft proofing It is very hard to estimate the monitor s characteristic without measuring, because there are no ways to know the monitor and environment conditions. The characterization contained in a monitor ICC profile is obtained using default factory values and a dark room. Thus, the information of the factory profile differs from the real output color of monitor. 8 In order to obtain a good estimate of the monitor characteristics and environment, we propose a method based on soft proofing. First, the input image or test chart is printed by direct printer driver, as shown in Figure 7. And the RGB values of input image are converted to CIEXYZ values with printer s ICC profile. Next, the CIEXYZ values are reconverted into the new RGB values with monitor s ICC profile. The color of output image by conversion process with ICC profiles represents the color of printed image on the monitor. But, if monitor ICC profile is not same with characteristic of real monitor, the color difference between printed color and converted color on the monitor is still existed. To reduce the color difference and match two images, ICC profile of monitor is corrected by multiplying gain values to the RGB channel and modifying the gain value of GOG model, because the change of monitor characteristic by varying user mode of monitor can be estimated by two sort of gain parameter. If the color matching is completed, the monitor s ICC profile is then estimated. Printer driver Input RGB CIEXYZ Printer s ICC profile Monitor s ICC profile Printed image Matching Output RGB Printer Monitor Figure 7. Estimation of monitor ICC profile using color matching by soft proofing. SPIE-IS&T/ Vol Q-6

7 3.2 Color conversion After estimating monitor s ICC profile by soft proofing and color matching process, the output image is converted as shown in Figure 8. Input RGB values are converted to CIEXYZ values using the estimated monitor s ICC profile, and the CIEXYZ values are reconverted into new RGB values by printer s ICC profile. When the corrected RGB values are printed, the color becomes similar to the monitor s color. Input RGB CIEXYZ Estimated monitor s ICC profile Printer s ICC profile Output RGB Generation of output image Figure 8. Color conversion process with estimated monitor s ICC profile. 4. EXPERIMENTAL RESULT In experiments, the color temperature and contrast mode of monitor is changed to cool mode and 90 as a contrast mode. As shown in Figure 9(a), the image on monitor is bluish and has high contrast. However, the color of directly printed image, Figure 9(b), is more grayish or yellowish and the contrast is lower than the image on monitor. The converted and printed image, Figure 9(c), has similar color and contrast to the image on monitor, Figure 9(a). (a) (b) (c) Figure 9. (a) Captured image on the monitor, (b) printed image by direct printer driver, and (c) converted and printed image. 5. CONCLUSION A monitor s color performance can be changed by user settings and age of the monitor itself. It is hard to estimate the color of monitor by predefined data. Thus, the characteristic of monitor should be estimated to represent the color of monitor into the color of printer. We proposed a method to estimate the monitor characteristics by color matching based on soft proofing. The monitor characteristics are estimated using a printed reference image. Then a new monitor ICC SPIE-IS&T/ Vol Q-7

8 profile is created and used for output image. Our results show that the output from the monitor better matches the output from a printer in the same environment when using out method. ACKNOWLEDGEMENT This work was supported by the Korea Science and Engineering Foundation (KOSEF) grant funded by the Korea government (MEST) (No ) and this work was supported by the second phase of the Brain Korea 21 Program in REFERENCES [1] H. Zeng and M. Nielsen, Color Transformation Accuracy and Efficiency in ICC Color Management, IS&T Ninth Color Imaging Conf., (2001). [2] T. Kohler, The next Generation of Color Management System," Proc. IS&T/SIC 8th Color Imaging Conf., (2000). [3] C. S. Lee, Y. W. Park, S. J. Cho, and Y. H. Ha, Gamut mapping algorithm using lightness mapping and multiple anchor points for linear tone and maximum chroma reproduction, Journal of imaging Science and Technology, 45(3), (2001). [4] J. Morovic and M. R. Luo, Evaluating gamut mapping algorithms for universal applicability, Color Research & Application, 26(1), (2001). [5] [6] C. J. Edge, Correction techniques for soft proofing, US patent, US 7,209,147 B2 (2007). [7] C. J. Edge, Color Correction Using A Device-Dependent Display Profile, US patent, US 7,215,343 B2 (2007). [8] M. D. Fairchild, Color Appearance Models, Addison Wesley (1998). SPIE-IS&T/ Vol Q-8