Still more on the color illusion

February 18, 2018 • 7:45 am

Well, a reader showed in an email that that the hearts in the illusion below really were very slightly different colors due to the compression algorithms used in making the image. (However, those slight differences do not account for the striking perceptual effect).  The reader who demonstrated this anomaly won an autographed book from Matthew.

HOWEVER, reader Mel made his own illusion, so we know that here the colors of the squares really are identical. His notes:

To demonstrate the phenomenon a bit more cleanly I constructed the following image using a spreadsheet. I used very small cells (0.3 cm x 0.3 cm) and filled them with various colors and then took a screenshot. The image still shows the illusion and only three colors were used in constructing the image (magenta, orange, blue-green).

And another demonstration using magic markers. I think the efficacy of the lines in fooling viewers about the color has been shown. We’ll now leave this illusion behind and move on.

29 thoughts on “Still more on the color illusion”

      1. There are some rare forms of colourblindness that only see in scales of ‘grey’ (complete lack of functional cones), however their visual acuity is so bad that the lack of colour vision is but the least of their visual problems.

      2. I have deuteranopia, probably the most common form of colour deficiency (generally referred to as “colour blindness”), and can see the illusion very clearly. Monochromatic vision (greyscale black & white) is extremely rare. check out the Vision simulator to get an idea of how colour blindness affects sufferers.

    2. Up to 200 odd shades now, isn’t it? Unless the books, plays, movies and (probably) board games are getting a bit repetitive.

    1. You don’t need to offset by any vertical stripes. The “blue” square interrupts the orange lines, which is what makes it look blue, while the “green” one interrupts the magenta lines. So they could be directly adjacent if you want. I’m pretty sure it would look “the same” minus the separating distance.

      1. Actually, for me, that one doesn’t work. Same green, both sides of the shift.
        The spreadsheet version up-thread gives me a … I’m not sure of the name of the effect. When you look at the traditional drawing of a wire-frame cube-corner and your perception flips between being a view of the interior and of the exterior of the box. Well I get that same sort of flickering between the left box being green and the right one blue, and vice versa

        1. Necker Cube? For me the greens remain different. And no flicker or disorientation with the original.

          1. Necker cube – that’s the one.
            Lots of individual variation in optical systems. I remember being blown away by the prediction and then discovery of human tetrachromats.

          2. For me the two greens are very slightly different. I can kind of feel a tension in perception wanting to make the two sides into a uniform background?????
            So a differences in perceptual systems.

        2. Me too, I see it as the same green. Which now makes me want to look at images with a 2-stripe gap, 4-stripe gap, etc… to see at what point the illusion kicks in.

          4. Hmm, the top right image-looked at from a distance clearly is blue vs green but from close up it is easy to read it as a uniform green background. Oh well. Maybe I should read up on variability in human retinas of numbers of cells sensitive to different wavelengths.
            Thanks for all that computering.

          2. Thanks for posting those. I see slight difference in color at the “ends” of the squares in the first image, but practically none in the top/bottom image.

  3. I think it’s awesome how your brain automatically compensates for differences in ambient lighting and partially obscuring obstacles, to show you how each object contrasts with its environment. The fail here is much less impressive than the routine success. If only I could make my machine vision algorithms at my job work so well.

  4. On a slightly different but related (these illusions reside in the brain even though they are visual), NPR yesterday had a show on neuropsychology. One aspect was face recognition, and morphed faces (two faces morphed together by software). Apparently, if you morph yourself and someone well-known, you see yourself in the morph, others see the well-known person. Even stranger, if you anesthetize different sides of the brain and look at one of these morphed photos, if your left brain is anesthetized, you see yourself; but if your right brain is anesthetized, you (like everyone else) see the well-known person. Weird things go on in the brain.

    1. Yes, indeed. Possibly the same kind of experience as when you are expecting a friend to showup and you see them as you expect them to be- left brain? But sometimes catching a glimpse unexpectedlythe same friend can look completely different-especially if you have known them for a long time and they have aged or otherwise changed. The left brain goes with the built up expectation and the right brain presents a more up to date image. It can be quite startling. ????????? Or if you catch an unexpected image of yourself as a reflection in a store window..

      1. I don’t trust that left brain / right brain stuff – there is a lot of hype on the subject. It’s particularly difficult to untangle with respect to vision where some signals cross to the opposite brain hemisphere & some do not. And vision processing is a multi-step operation some of which is done before it hits the brain proper.

Comments are closed.