Study Of Color Reproduction Theory And Method For Digital Image

Various digital color image devices have been extensively utilized in daily life and work with the development of science and technology and the improvement of the living standard. However, when color images are reproduced from one digital device to another the color appearance of the reproduction images always disagree with the originals due to the color dealing mechanism and color gamut of the digital devices being different from each other, which seriously restrict their applications in the industries, such as textile'and printing, computer aid design, e-commerce, long-distance medical treatment and so on. The color management system is recognized as an effective way to resolve this problem. The key and fundamental techniques of the color management system were studied in depth in this dissertation. For the color management system based on colorimetry, the digital still cameras, liquid crystal displays and color printers were selected respectively as the typical representations of digital color image capturing devices, softcopy display devices and hardcopy output devices. The colorimetric characterization methods of these three kinds of devices and their gamut mapping algorithms were investigated separately. For the color management system based on spectral characteristics, the spectral reflectance reconstructing techniques were discussed in detail. Finally, the application instances were illustrated based on the research achievements of this study.For the colorimetric characterization of digital still cameras, the settings of the camera, the setup of the experiment and the illumination evenness and repeatability correction methods during the characterization process were analyzed. An adaptively polynomial model was proposed for the colorimetric characterization of the cameras, in which the affecting factors of the characterization accuracy were investigated in detail, including the nonlinear transformation, the number of local training samples, the number of polynomial items and sample medium. For the colorimetric characterization of liquid crystal displays, the two important assumptions, chromaticity constancy and channel additivity, were firstly introduced, and then the characterization principle based on these two assumptions were analyzed with the typical Gain-Offset-Gamma (GOG) model as an example. A professional liquid crystal display, which approximately satisfies these two assumptions, was colorimetrically characterized using GOG model to illustrate the characterization method and verify the characterization accuracy. Whereafter, for the commonly commercial liquid crystal displays that do not satisfy the two assumptions, a piecewise partition model (PP model) was proposed in this study for the colorimetric characterization of these displays. An experiment was carried out to compare the PP model with other colorimetric characterization models; the results indicated that the PP model performed best for the case of less training samples. For the colorimetric characterization of printers, which belong to the subtractive color mixture devices with strong nonlinear relationships between the driving signals and the corresponding colorimetric values, the cube interpolation and iterative cube subdivision models based on look-up table were adopted for the forward and inverse colorimetric characterization respectively, while the inverse model was further improved to promote the transformation efficiency. The experiment results illustrated that both the forward and the improved inverse colorimetric characterization models were of high color prediction accuracy.The color gamut boundary description method and gamut mapping algorithms were discussed and summarized due to their importance in the modern color management system. Afterwards, two new gamut mapping algorithms were proposed in this study, namely, the adaptively spatial color gamut mapping algorithm (ASCGMA) and image dependent gamut mapping algorithm based on Gaussian function compression (GGMA). In ASCGMA, the compression degree of the out-of-reproduction-gamut color is not only relating to the position of the color in the color space, but also depending on the neighborhood of the color to be mapped. In GGMA, the CIE color space was divided into many subspaces, and for the subspaces with no color in the image being out of the target device gamut, the colors in those subspaces were not compressed; otherwise, a modified Gaussian function was designed to map the colors in the corresponding subspaces. Both ASCGMA and GGMA were compared with HPMINDE and SGCK recommended by CIE according to the CIE guideline, and the results of the psychophysical experiment based on the pair comparison technique indicated that both the two proposed new gamut mapping algorithms outperformed HPMINDE and SGCK as a whole.The spectral reflectance reconstruction of surface color is the sticking point in the spectral color management system. The principal component analysis (PCA) is widely used to reconstruct the spectral reflectance. However, the estimated spectral accuracy is low when using only one set of three principal components for the three-channel digital color image devices. In this study, the spectral space was firstly divided into 11 subgroups, and the principal components were calculated for individual subgroups, and then the principal components were further extended from three to nine. For each target sample, the CIEXYZ tristimulus values and the nine principal components of the corresponding subgroup samples were based to reconstruct the spectral reflectance. The experimental results illustrated that this proposed method was quite accurate and outperformed other related methods. In order to further investigate the validity of the research production in this study, two instances were carried out, i.e. the cross-media color image reproduction from a liquid crystal display to a color printer and the spectral reflectance reconstruction based on the printer driving signals. For the first instance, four standard images were reproduced from display to printer through the modern color management system based on the research achievements in this study, and the mapped images were compared with the printed images under the default settings of the printer. The psychophysical experiment results showed that all the hardcopy images processed by the color management system were more similar to the original images with comparison to the printed images by default. For the instance of spectral reconstruction, the spectral reflectance of the testing color patches were reconstructed based on the printer driving signals through the proposed method in this study, the reconstructed spectral reflectances were compared with those measured data, and the results indicated that the proposed method could accurately reconstruct the spectral reflectance.Finally, the main contents and innovations of this dissertation were summarized, and the perspectives of the future study were also forecasted.