The Relationship Between Distortion and Image Quality: Sampling Procedures and Accuracy

摘要

The time variability in subjective assessments of image quality is considered as well as its effect on the evaluation of the mean opinion in double stratified samples of observers viewing still images. Attention is focused on a sampling procedure which avoids the factors that might introduce a bias in the function used in the least squares fit of experimental data. A method to evaluate the accuracy of the fitted function and to verify the validity of the additional law of impairments is also presented.