Quality analysis for coded images with loss

Abstract

A considerable compression rate can only be achieved by means of algorithms with loss, which means that it is not possible to recover the exact original image. This loss of information may have a direct relation to the loss of quality, as well as cause problems related to reliability depending on the area of the application. The main question to answer is: How do we decide that an image compressed with loss is suitable to be used for a given application? This question is answered by defining a subjective quality assessment and then relating it to objective values in order to find - by means of a statistic analysis of the data - a relationship between quality and compression ratio. In particular, the incidence of the type of histogram and its relation to subjective loss and compression ratio are studied. This paper is related to the Magister Thesis “Análisis del Error en Algoritmos de Transmisión de Imágenes Comprimidas con Pérdida” ("Error Analysis in Transmission Algorithms of Images Compressed with Loss") by Lic. Ramón