Research On Key Technologies For Color Printer Characterization

Image colors require multiple conversions during the workflow of color image output system. However, there exist great differences in imaging mechanism among various kinds of image devices, even the same type of devices will vary as well. As a result, the color management needs to be carried out in the output system. In ICC-based color management, the device-independent color space of CIE L*a*b*or CIE XYZ is chosen as the Profile Connection Space (PCS). Colors of different devices can be converted through PCS merely by establishing the relationship between device color space and PCS in advance, thus achieving a device or platform-cross consistent color reproduction.This dissertation selects color printer as the research device. By analyzing the key technologies involved in printer characterization, it focuses on the study and innovation on the respect of printer characterization method, gamut mapping algorithm and the selection of characterization sample set, and then designed and implemented a printer color management system. The research results achieved a significant reproduction precision for image colors.Firstly, the common methods of printer characterization are reviewed and an improved method is proposed by analyzing their advantages and disadvantages. The common characterization methods include model-based method, polynomial regression, multi-dimensional look-up table method (MLUT), and artificial Neural Network, where the MLUT method has the highest accuracy. The experimental analysis indicated that the printer inks have nonlinearity themselves, which will lead to more serious nonlinearity and color inconsistency after ink mixture and overprint. So the printer ink calibration before characterization can reduce the nonlinear strength effectively. For the traditional MLUT method is unable to ensure a minimum global error, a novel characterization method based on least square method is presented. Experiments results showed the precision of the ICC profile generated by this method was superior to that of traditional MLUT method.Secondly, in terms of the existing problems in current gamut mapping algorithms, three improved algorithms are put forward. ICC specification recommends SGCK method for perceptual intent and HPMINDE method for colorimetric intent, without mentioning the proper gamut mapping method for saturation intent. This thesis proposed a saturation-based gamut mapping algorithm according to the properties of SGCK and HPMINDE. This algorithm can reserve the saturation of colors as much as possible and keep the relationships between colors as well. It is applicable to the gamut mapping for computer graphics, such as pie graphs and maps. Since the existing gamut compression methods fail to consider the spatial characteristics of the image, a gamut compression algorithm based on image spatial characteristics is proposed. This method takes into account spatial characteristics in addition to the traditional SGCK, and achieved the goal that the contrast invariances of overall colors are guaranteed and the spatial characteristics are also reserved. In current color management workflow, the end users must select the rendering intent according to the image features and the relationship between image and reproduction gamut. Concerning this issue, an adaptive gamut mapping algorithm based on image is put forward. The image is first classified with image features, and then the optimal rendering intent can be chosen automatically on the basis of image class and the relationship between image and reproduction gamut, thus improving the current color management workflow.Afterwards, with respect to the selection of characterization sample set for color printer, a sampling optimization method based on color difference analysis was presented. The core idea of this method is that when using ICC profile to predict the colorimetric values of colors in the region of CMYK color space, if there are great color differences between predicted values and actual outputs, it indicates the sample number of characterization sample set in this area lacks and more samples need to be added to increase the corresponding accuracy. The sample set obtained with the above method comprehensively takes the nonlinearity of different areas in printer color space into account and achieves a significant reduction in the number of samples.Finally, on the basis of the above research, an ICC-based printer color management system is designed and implemented. It is composed of generation module of ICC printer profile, printing module and evaluation module. Experiment results showed this printer color management system is able to obtain a great accuracy of color reproduction and can satisfy the requirement of color control for printers.