The quality analysis of image reproduction

The objective of image reproduction is, as used here, to reproduce visual information that is available in the original image. Undoubtedly, the colors of the individual pixels contain a huge amount of visual information and therefore the quality of the image reproduction is typically studied in terms of the quality of the color reproduction determined for each image pixel. It is, however, true that not only the colors of the individual pixels but also the spatial structures formed by the color fluctuations across the image are very important sources of visual information. Because only minor errors in the color reproduction may cause significant errors in or even vanishing of the spatial structures of an image, they both—the reproduction of the color and the reproduction of the spatial structures—have to be analyzed to determine the quality of the image reproduction. There are commonly accepted methods for the quality analysis of color reproduction. A quality analysis for the reproduction of the structures is developed in this study. Three metrics, selected on the basis of the monocular sources of visual information, are used to describe the spatial structures of the original and the reproduced images. The metrics—the edge contrast, the size of detail and the evenness contrast—are analyzed by using the methods of binary and gray morphology. The use of the metrics in the quality analysis of the image reproduction is illustrated by computer simulations.