[Gathnet] Optical illusion of motion

Jonathan Gathman jonathan at stl.gathman.org
Fri Jul 2 19:05:46 EDT 2004


Don't all optical illusions present something that is so close to what 
our mind expects, that it tends to adjust the picture?  After all, we 
have the nerve connections in the back of our eye, slightly the outside, 
where there aren't any actual receptors.  We never see this "hole", but 
our minds fill it in based on context.  (try holding a pencil with an 
eraser down and to the left (given your left eye) of the center.  Move 
it around, and see if you can make the eraser part "disappear".)

I think our eye tries to line up what looks like little cylinders, the 
ends represented by the blue and yellow ovals, but the cylinders are not 
placed where the mind expects, so every eye movement changes the 
perspective slightly.

Of course, all computer screens are optical illusions as well, as they 
are really flashing pixels, refreshed at very fast intervals.  To 
disprove this, I printed it out.  While the colors are not as vibrant 
with my printer, they still seemed to move, so we can discount the 
computer monitor element...

Thanks Scott, this is really cool!


SKM wrote:
> http://www.cs.toronto.edu/~moraes/illusion.html
> 
> I do not know how it works, but I will bet that it is related in part to 
> the color of the ovals. If you think of this as a collection of segments 
> consisting of two ovals (blue and yellow) with a connecting black 
> biconcave lens, all the circles appear to rotate "away from yellow" and 
> "towards the blue". If you convert the picture to grayscale, the 
> illusion is still present but less strong. If you rotate the image 180 
> degrees, the blue and yellow circles are on opposite sides, but the 
> rules just mentioned apply. If you make it a "negative" (inverting all 
> colors) these rules still apply. You can see this things yourself by 
> saving the picture from your browser to your desktop and then 
> manipulating it with a utility such as Irfan view.
> 
> There is an explanation that I do not understand at 
> http://psych.upenn.edu/backuslab/vss/vss2004/backus2004.html
> Scott
> 
> _______________________________________________
> Gathnet mailing list
> Gathnet at gathman.org
> http://bmsi.com/mailman/listinfo/gathnet
> 

-- 
Jonathan Gathman
"Think, Speak and Do Well"
Urquhart Family Crest





More information about the Gathnet mailing list