Abstract: Color image quality prediction models for two typical documents used as input for color copying machines have been developed to relate subjective image quality ratings to physical image quality metrics using stepwise multiple regression analysis. The typical documents consist of colored map and portrait images. The models were consistent with technical knowledge and achieved high correlation between predicted ratings and measured subjective image quality ratings. By utilizing the models, subjective color image quality can be measured by instrumental measurements, and also color imaging system of preferred image quality can be designed by physical image quality metrics, and this leads to effective image quality design. !9
展开▼
