Physical Criteria for the Study of Compression with Losses in Digital Medical Images: a Review of the Subject

Abstract

The evaluation and description of image quality after applying lossy compression is described. It remains an open research issue in the sense of finding an objective criterion to stop lossy compression at a maximum level of compression with a minimum distortion. This value is found by implementing a set of quantitative bivariate measures such as the Mean Square Error (MSE), the Peak Signal to Noise Ratio (PSNR), the Structural Simmilarity index (SSIM index) and the Spectral Distance among others. A review is also done about the physical meaning of prediction algorithms for quality comparison such as the Hotelling Observer (HO), the channelized forms (CHO and LGHO), and the Non-Prewhitened Matched Filter (NPWE). Finally, the subjective evaluation of image quality is investigated through ROC analysis based on diagnostic performance curves.