Picture quality assessment model
T. Weyers, Johan Schalkwyk · Electronics Letters · 1993
The derivation of a model that simulates human opinion on distortions encountered in stationary transform coded video images is discussed. Regression of subjectively relevant errors measured in the compressed images to a mean opinion score obtained from a group of observers was employed. Results show a standard error of 11.44% for the average human observer's prediction of the mean opinion score, and the standard error of the prediction model is 10.82%.