A new psychophysical experimental method is proposed for estimating image quality. The method comprises two steps; the first step is a 'categorical session' and the second step is a 'triplet comparison session' which is newly developed for this purpose. The triplet comparison method is developed, in order to improve the assessment accuracy and repeatability, without imposing excess observer stress during the visual assessment. Moreover, since the triplet comparison inevitably reduces the number of comparisons, experimenter can design psychophysical experiments using a higher number of samples, compared to the paired comparison method. A series of psychophysical experiments were conducted and performed. It is found that the proposed method gives reliable and stable results so far examined.
|