Research On Color Gamut Mapping And Its Evaluation

The cross-media high fidelity reproduction of color image has become a research emphasis in the color reproduction field with the extensive application of digital media,and in order to avoid color distortions in the cross-media transmission color gamut mapping algorithm(GMA)is usually adopted which is also an critical stage of color management system.In this paper,several key and ffundamental techniques of GMA were studied intensively focusing on the theory and evaluation method of GMA to improve the current GMAs' performance.The image quality assessment models were studied to testify the feasibility in the evaluation of GMA as well.Color gamut mapping algothims were improved from the perspective of color gamut visualization.The accurate of gamut boundary description can benefit the out-of-gamut determination of images and eliminate the uncontrolled clipping or compression of GMAs.The gamut boundary description(GBD)algorithms were analyzed in detail,and the comparision between the segment maxima method and the GBD algorithm based on irregular segmentation was conducted,the results indicated that the gamut boundary descriptors using the irregular segmentaion based GBD algorithm were distributed uniformly to some extent among the gamut surface,then a new GMA was proposed based on the irregular segmentation GBD algorithm.In the example of this new GMA in this paper,CIELAB color space was selected as the mapping space and minimum color difference was selected as the mapping criterion.An GMA evaluation experiment was carried out to compare the new GMA with the HPMinDE GMA recommended by CIE,the results indicated that the quality of mapped images was improved.In the aspect of GMA evaluation method,several image quality assessment models(IQMs)including S-CIELAB,SSIM,MSSIM and CID were analyzed in depth,hite rate method was adopted to verify the performance of these IQMs using Empa gamut-mapping database,the results indicated that IQMs can be successfully used to evaluate the performance of GMAs,but the performance of IQMs still need to be improved.