Image quality analysis of the reproductions of black and white photographs obtained from a desktop publishing system

This research project was directed to characterize the variables that govern the black and white reproduction process in a desktop publishing enviroment (DTP), which in this case is formed by: a Macintosh II computer, a Mycrotek 300A scanner as the inputing device and a LaserWriter as the outputting device. The specific goal of this research was to find the ideal settings of cell size and resolution for each type of black and white photographic image that would produce the best reproduction possible from the system. Two different experiments were performed, the first one using a gray scale and a resolution target as originals, showed the influence of the variables over the tone reproduction curve and the resolution of the system. The second test used three different images as originals, and their corresponding reproductions were rated by a selected group of judges, using the pair comparison technology for the evaluation. In addition to obtaining the best settings for each image, an analysis of the application of information theory to image evaluation was performed. The results found that the actual mathematical model needs to incorporate other factors such as the visual response of a human observer, to be considered as a valuable tool for quality assesment of a reproduction system