If you, like me, are one of today's lucky ten thousand[1], here's some context:
Kokichi Suhihara is an engineering professor who designs 3D objects that create optical illusions, including their mirrored images. Here's his English home page:
So we're looking at a photo of an actual 3D object and its mirrored reflection, which should obviously appear incorrect since, well, the reflection is not apparently mirrored but rather facing the same way as the real object. I think the key to understanding what we're seeing is that the non-mirrored image is also not as it appears. We'd need to see it rotated to understand the actual shape.
Here's a short YouTube video of his "Ambiguous Garage Roof" that I've now watched several times and it keeps blowing my mind:
Funny story: In 1984, my mom, a Computational Geometry researcher, saw a proof [1] in some IEEE journal that she was annoyed with. It seemed way too general and was too easy to disprove, so she did [2], in a subsequent publication, with the author's name in the title.
She later told me he courteously publicized a correction, credited her, and would mail her pre-prints for his papers since. Earlier this year I looked up the claim and counterclaim: Turns out it's Kokichi Sugihara.
I guess in the 80s well-actuallying in Mathematics journals was good form. :)
What's amazing about the Ambiguous Garage is that even after I've seen how the illusion is done, my brain keeps getting fooled. It's one of those things where it's simply impossible not to "see" the illusion even after you understand it.
It reminds me of the Penn and Teller red ball routine where they tell you how it works before they perform it and your brain (well, mine anyway) still won't accept it.
Searching on YouTube doesn't seem to bring anything about the one linked here, am I putting in the wrong term? I'd love to know how this was done, it's messing with my head.
Oh, this is really interesting, he goes into detail on how difficult it can be to print them, too, based on the fact that you have to make the parts moveable and a lot of them are somewhat small, too. Thank you!
Thanks. I meant it literally and figuratively. Literally, drawing shadows of optical illusions generally makes them stop being optical illusions. Figuratively, all illusory experiences (magic tricks, scams, psychotic delusions) tend to have some mundane inconsistency which is likely to go unnoticed, but if you know to look for it the illusion evaporates. Illusions don't have shadows.
I've printed it and physically fiddling with it made it much more clear how it worked.
I haven't seen a print for that dog, though.
In theory, I think you could use Sugihara's approach to automatically generate such a model for most closed curves. That'd be a cool little program that I don't know how to write.
Kokichi Suhihara is an engineering professor who designs 3D objects that create optical illusions, including their mirrored images. Here's his English home page:
http://www.isc.meiji.ac.jp/~kokichis/Welcomee.html
So we're looking at a photo of an actual 3D object and its mirrored reflection, which should obviously appear incorrect since, well, the reflection is not apparently mirrored but rather facing the same way as the real object. I think the key to understanding what we're seeing is that the non-mirrored image is also not as it appears. We'd need to see it rotated to understand the actual shape.
Here's a short YouTube video of his "Ambiguous Garage Roof" that I've now watched several times and it keeps blowing my mind:
https://youtu.be/KtA6u1HIqbg
[1] https://xkcd.com/1053/