A number of quality measures were evaluated for grayscale image compression. These measures are bivariate, exploiting differences between corresponding pixels in original and degraded images. While some numerical measures correlate well with observer responses for a given compression technique, they are not reliable for evaluation across different techniques. The two graphical measures, histograms and Hosaka plots, can appropriately specify both the amount and type of degradation in reconstructed images. The study evaluated the performance of various image quality measures, compression techniques, and test images. The compression techniques included JPEG, EPIC, RLPQ, and SLPQ. The test images were Lenna, Fingerprint, and Hurricane Gilbert. The spatial frequency of an image was defined based on pixel differences. The performance of quality measures was assessed by comparing numerical objective measures with subjective evaluations. The Pearson product-moment correlation coefficient (r) was used to measure the linear relationship between the two. The results showed that NMSE(HVS) was the best measure for all test images. Other measures, such as NAE and L2, did not show significant improvement when incorporating the human visual system (HVS) model. Histograms and Hosaka plots were found to be effective graphical measures. Histograms can indicate the type of degradation, such as blockiness, blurriness, and fuzziness. Hosaka plots provide a detailed description of the type and amount of degradation by plotting feature errors in polar coordinates. The results showed that JPEG preserved high-frequency components, while RLPQ induced uniform blockiness. The study concluded that while numerical measures can reliably specify the magnitude of degradation for a given compression technique, they are not suitable for evaluation across different techniques. Graphical measures, such as histograms and Hosaka plots, are more effective in specifying the type of degradation. A combination of numerical and graphical measures may be more useful in judging image quality. Further research is needed to develop new graphical measures with superior judgment capabilities.A number of quality measures were evaluated for grayscale image compression. These measures are bivariate, exploiting differences between corresponding pixels in original and degraded images. While some numerical measures correlate well with observer responses for a given compression technique, they are not reliable for evaluation across different techniques. The two graphical measures, histograms and Hosaka plots, can appropriately specify both the amount and type of degradation in reconstructed images. The study evaluated the performance of various image quality measures, compression techniques, and test images. The compression techniques included JPEG, EPIC, RLPQ, and SLPQ. The test images were Lenna, Fingerprint, and Hurricane Gilbert. The spatial frequency of an image was defined based on pixel differences. The performance of quality measures was assessed by comparing numerical objective measures with subjective evaluations. The Pearson product-moment correlation coefficient (r) was used to measure the linear relationship between the two. The results showed that NMSE(HVS) was the best measure for all test images. Other measures, such as NAE and L2, did not show significant improvement when incorporating the human visual system (HVS) model. Histograms and Hosaka plots were found to be effective graphical measures. Histograms can indicate the type of degradation, such as blockiness, blurriness, and fuzziness. Hosaka plots provide a detailed description of the type and amount of degradation by plotting feature errors in polar coordinates. The results showed that JPEG preserved high-frequency components, while RLPQ induced uniform blockiness. The study concluded that while numerical measures can reliably specify the magnitude of degradation for a given compression technique, they are not suitable for evaluation across different techniques. Graphical measures, such as histograms and Hosaka plots, are more effective in specifying the type of degradation. A combination of numerical and graphical measures may be more useful in judging image quality. Further research is needed to develop new graphical measures with superior judgment capabilities.