| Ahmet M. Eskicioglu, Paul S. Fisher, and Siyuan Chen
The paper evaluates several quality measures for grayscale image compression, focusing on their performance and reliability across different compression techniques. The authors use a set of experiments to assess the effectiveness of these measures, including numerical and graphical methods such as histograms and Hosaka plots. They find that while some numerical measures correlate well with subjective evaluations for a specific technique, they are not reliable for comparing across different techniques. In contrast, the graphical measures can specify both the amount and type of degradation in reconstructed images. Specifically, the NMSE(HVS) measure is the best among the numerical measures, and Hosaka plots effectively illustrate the types of degradation caused by different compression techniques. The study concludes that a combination of numerical and graphical measures may be more useful for judging image quality, and further research is needed to develop new graphical measures with enhanced judgment capabilities.The paper evaluates several quality measures for grayscale image compression, focusing on their performance and reliability across different compression techniques. The authors use a set of experiments to assess the effectiveness of these measures, including numerical and graphical methods such as histograms and Hosaka plots. They find that while some numerical measures correlate well with subjective evaluations for a specific technique, they are not reliable for comparing across different techniques. In contrast, the graphical measures can specify both the amount and type of degradation in reconstructed images. Specifically, the NMSE(HVS) measure is the best among the numerical measures, and Hosaka plots effectively illustrate the types of degradation caused by different compression techniques. The study concludes that a combination of numerical and graphical measures may be more useful for judging image quality, and further research is needed to develop new graphical measures with enhanced judgment capabilities.