It's very cool - especially the connection to object detection - but as the video demonstrates, it's only functional from a limited field of view. If you see the area from a different/unexpected angle, the camo stands right out. So it's great for fooling stationary sentries or cameras, but it's no invisible cloak. Yet.
Its only a matter of time before this shows up on a Japanese game show. The host will hide in plain sight, then jump out and scare the "victims".
I was unclear from that video: did they actually try this in the real world or are those just computer-generated simulations?
A little narration would go a long way in helping me to understand (some) of the use cases.
I was stumped by the tree ones. Seriously. I couldn't see the "thing" in a bunch of them.
I could imagine a future where armies rapidly produce single use uniforms (or just clean suit like covers) super customized to the environ of the target ... and the squadron opens their box to find navy-blocked blues while they're in afghanistan.
that's because it was behind you the entire time...
This was one of the things that I disliked about the Hunger Games movie - going to great lengths to exactly reproduce your background is definitely secondary to finding a good place to hide and blurring your outline.
This kind of individualized camouflage would probably be useful for hiding equipment though, especially if you were only likely to see it from one angle (from the air or a nearby road, perhaps). During wwii they disguised entire factories as villages, so it could have some potential as more convincing camouflage. I wonder how well it works with different lighting and shifting shadows?
I had to watch twice to even see those things. Holy crap, it's that cloak Harry Potter had.
I'm not sure I get what the new part is.
You can do it the other way, as well. Instead of putting the image of the background onto foreground objects you can put the image of foreground objects onto the background. The whole area has been useful in visual effects for some time. And chalk drawings on sidewalks.
But, those uniforms totally work:
What, you couldn't see the manhole cover? Perhaps because it was disguised to look exactly like a manhole cover.
But seriously, the guy is cheating by using computed interpolation to show "movement" about the scene. I'm certain his little camo boxes would be obvious if he shot actual video from a moving point of view.
By taking still shots from the exact viewpoints for which the camo panels were computed, and then including only those vewipoints in the computed interpolation, he's making it look better than it really is.
Ha. I just read the linked article. They are virtual boxes, generated by the computer and superimposed on the scenes.
So these guys have done even less than I thought before. They should at least be required to print out their experimental textures, glue them to the sides of a box, and demonstrate them under ambient lighting in the real world. I'm calling their research project BS, probably aimed at getting a BS degree.
This topic was automatically closed after 5 days. New replies are no longer allowed.