DISCRIMINATION OF CHANGES IN THE 2ND-ORDER STATISTICS OF NATURAL AND SYNTHETIC-IMAGES

It has been suggested that the second-order statistics of different na tural images are all remarkably similar and that neurones and channels in the visual system may exploit this similarity. We have measured th e ability of human observers to discriminate changes in these statisti cs using different natural and synthetic stimulus images and have foun d that the dependence of their discrimination thresholds upon the refe rence second-order statistics is similar in form, for both kinds of st imuli. However, there is some variety in the magnitudes of the thresho lds for the natural stimulus images; in fact, the second-order statist ics of different natural images are more diverse than previously sugge sted. The discrimination task can be modelled as the discrimination of changes in local contrast within restricted spatial frequency bands a nd is similar to the discrimination of blur.