Study on Methods of Image Quality Evaluation Based on the Color-Difference Formula
|School||Beijing Institute of Printing|
|Course||Signal and Information Processing|
|Keywords||Digital image Subjective evaluation color-difference formula Psychophysical experiment|
With the growing of imaging technology application, there are higher and higher requirements for images display in all industries referring to color images. By now, there is no method for evaluating image color-difference specially. The traditional methods are mainly not aimed at the image color. At the same time, their calculations are complex and not based on the human visual. Recently, Researchers in related fields are actively study the methods to predict image color-difference sense based on color-difference formula. In order to find a way, which is fast and effective, to evaluation quantitatively image color-difference.Color-difference evaluation is quite important for the color quality control of industrial products, and color-difference formulas are scientific and effective methods to evaluate the color difference between two uniform color samples. As the image is a combination of non-uniform and complex color samples, which contains a large number of pixels with different colors. The image color difference cannot be measured directly by instrument, and the calculated results from original color-difference formulas are not stand for the visual color difference sense. Thus, there is the demand for establishing the link between objective and subjective evaluation value based on visual experimental data and calculation results. Results from different color difference formulas are different. Therefore, the ability assessment of different color-difference formula that predicts the image color-difference sense is very urgent and imperative research, and there is important theoretical significance and practical value.A large number of digital images and subjective experimental samples are produced, not only for much convince, but also for the establishment of relationship between subjective and objective color-difference. Subjective experiment of image color-difference levels evaluation is set up to establish the base data of image color-difference calculation. The ability of some typical color-difference formulas that evaluates image color-difference sense of is tested, including CIELAB、CMC、 DE2000 and S-CIELAB model. The performance of different formulas in both original and revised form is compared. In addition, the sense of image color-difference threshold by CIELAB calculation is got through image color-difference threshold experiment. At the same time, the visual experimental data are also collected by CIE organization for the establishment of color image evaluation data base. The research that evaluates the ability of different color-different formula to predict image color-difference provides data and idea for evaluating complex image quality, and lays the foundation for the color-difference theoretical research.