The Craik-O'Brien-Cornsweet illusion reference article from the English Wikipedia on 24-Apr-2004 (provided by Fixed Reference: snapshots of Wikipedia from wikipedia.org)

Craik-O'Brien-Cornsweet illusion

The Craik-O'Brien-Cornsweet illusion, also known as the Craik-Cornsweet illusion and the Cornsweet illusion, was discovered by Tom Cornsweet in the late sixties.

Adjust your browser so that you are seeing all of the picture below, and describe to yourself what you see.



If the entire region to the right of the "edge" looks slightly lighter than the area to the left of the edge, then you are experiencing the illusion. Of course, as those familiar with the presentation of optical illusions may have guessed, the brightness of the areas is exactly the same, as we can see if we black out the region containing the "edge:"



The following picture shows the actual distribution of luminance in the picture, and the typical perception of luminance:



The illusion is a little subtle. It is similar to the familiar phenomenon of simultaneous contrast and Mach band, but differs from it in two important respects.

• In most contrast effects, such as Mach bands, the effect is seen only on areas that are close to the the intensity gradient. In Craik-O'Brien-Cornsweet illusion, however, a very small area (the central "edge") affects the perception of entire large areas, portions of which are distant from the "edge."
• In the Craik-O'Brien-Cornsweet illusion, the region adjacent to the light part of the "edge" appears lighter, and the region adjacent to the dark part of the edge appears darker, just the opposite of the usual contrast effects.

The images above show the usual presentation of the effect. A far more convincing and dramatic version of the effect can be seen in Purves, Lotto, and Nundy (2002) where it is presented within a quasi-realistic image of solid, illuminated objects. These writers give a (teleological) explanation of this and other illusions, in which the visual system and brain are posited to perform a kind of Bayesian analysis of the likelihood of various real scenes producing the observed brightness distributions. In their words, "[perception] accords not with the features of the retinal stimulus or the properties of the underlying objects, but with what the same or similar stimuli have typically signified" in the past.

External links

• Purves, Dale, R. Beau Lotto, Surajit Nundy, 'Why We See What We Do'", American Scientist,'' 90(3):236-242 (entire article)
• Dramatic image from Purves, Lotto, and Nundy showing a much more convincing presentation of the Cornsweet illusion

This is the "Craik-O'Brien-Cornsweet illusion" reference article from the English Wikipedia. All text is available under the terms of the GNU Free Documentation License. See also our Disclaimer.