| As the society has been arounded by the information, digitization and network，the monitor which people usually get information through it gradually become moreand more poplulor. Color display quality has been one of the greatest issues of thevarious display manufacturers and consumers. However, CRT, LCD, LED largescreen, OLED, etc, has different levels issues in image cross-media color reproductionbecause every kind of it has different mechanisms to show colors. In this paper, acomprehensive study has been given about a serious of key technologies in thedisplay color management system.Display chroma characterization. First, this article describes four displaycharacteristics of color displays. Then, three kinds of method, Look-up table,polynomial and neural network, based on target color have been proposed. For thecharacterization of CRT and LCD monitors, this paper presents the mathematicalmodel of the LCD and CRT display characteristics. The PR655Colorimeter has beenused to characterize a LCD display in darkroom, then characterization data and curvefitting results have been given. This paper first introduces the LED light-emittingcharacteristics and spectral shift for LED large screen. Most of the LED display usesPWM dimming to solve the problem. This paper analyses the PWM driving principleand establishes the Characterization model of the LED display, which have been usedto characterize an indoor LED cabinet, then the experimental data and fitting resultshave been given.In order to solve the problem that it is hard to quickly and accurately calculatethe coordinate of intersecting point between the mapping line and the gamut boundary,an iterative calculation method based on the improved CORDIC algorithm toapproach the intersection has been proposed. The method using the characteristics of CIE LUV color space can approach the point along the mapping line. It can save a lotof storage resources and computing time because it has no boundary search andinterpolation calculation process. In addition, it has high calculation accuracy andwide applicability. This paper analyzes the algorithm of the calculation principle,precision and speed. With the LED Display, for example, it is made a boundarycalculation and the error analysis in D65light source. Experimental results show: Thefitting boundary is very smooth after12iterations and maximum color difference isonly0.16. The total computing time spent by calculating500intersections of mappinglines is about1s. Compared with the interpolation method, the maximum colordifference decreases by2.15， computing time reduced from10s to1s andcalculation speed increased nearly10times.In order to solve the problem that "Device to device" gamut mapping algorithmcan not reproduce the details of the final image，an "Image to the device" algorithmbased on MRF-MAP model has been proposed. Two parts devided by the algorithm torestrain the final reproduced image: The conditional probability of the image，whichcommonly uses "device to device" gamut mapping algorithm, restricts the gamut offinal reproduced image; The priori probability of the image, which commonly usesMRF model, extracts local details of original image. Finally, the maximum aposteriori probability estimation algorithm is used to handle this optimization problem.Advantage of the proposed algorithm is that priori information and local details of theimage could been integrated into the algorithm and the final reproduced image isalways exist and unique. Image mapped to the LCD display which captured by digitalcameras is used to evaluation result of the proposed algorithm with the gamutmapping algorithm recommended by CIE. Evaluation criteria is the z-score. |
PDF Full Text Request