[TheTarquin]

What's even more interesting is how much of the image you can remove while still maintaining the illusion. Block out almost all the image except for squares A and B and just a VERY little bit of space around them. The two squares still look different colors.

The brain can mess with perception in pretty significant ways based on very little information.

[joeyo]

The brain can mess with perception in pretty significant ways based on very little information.

Indeed. Which isn't too surprising since perception constructed by the brain. :) I think that the fact that you can't "will" yourself to see the colors how they "really are" suggests something the brain is taking some sort of computational shortcut to allow us to process light and shadows efficiently--one that is almost never wrong in the environment that we evolved in--that we can violate the assumptions of in the illusion.

[TheTarquin]

Which isn't too surprising since perception constructed by the brain. :)

Touché. I really should have said something like "the subconscious can mess..." or "the lizard brain can mess...".

[J_McQuade]

I tried covering up bits with my hand, with bits of paper, and staring at it with varying intensities - I still didn't believe it until I loaded it up in Gimp and checked the colour values.

It seems our perception of colour could be like our perception of temperature - comparative, rather than 'quantitative' in any way. Anyone here know of any studies into this?

[evanrmurphy]

I even got suspicious that the right image on the Proof page might have been doctored, but it does pass the Gimp test. :)

[JoeAltmaier]

Make a little circle with your index fingers and thumbs, put it over each square, that breaks the spell.

[TheTarquin]

An interesting idea. This would put it in the same class as a remarkably large number of emergent "senses" like our sense of distance or of weight, which are both effected by ancillary information.