Optical Camouflage

Mike Siekkinen writes "This optical camouflage project is pretty interesting. It contains three videos demonstrating it in action. Basically they overlay a video projection of what the background behind the object to be camouflaged looks like. So if you were standing in front of a book shelf an image of the portion of the bookshelf you are blocking would be projected on to you. The results are probably better than you would expect."

Comments

It's still visible

[Anonymous Coward]

And it requires a projector to paint the image.

So it won't work in a misty or foggy scenario like a jungle or in winter.

It might work in the desert, but so does netting.

Back to the drawing board!

Re:holy sh*t

[pardey]

It certainly looks good in the videos, but I wonder how well it does if you're off-axis? If I'm reading the diagram right, it looks like you's start to get parallax error pretty quickly.

Various translations...

[sporty]

Since we all don't speak the native languages of English or Japanesse, here are the translated versiosn from....

English to Japanesse

and...

Japanesse to English

We can't let those scientists have the fun of putting up bilingual pages up. After all, what's science without bad translations, eh?

Still can't figure out how it works...

[alphaseven]

Saw this on fark and boingboing a few days ago.

A number of posters on those site feel it's no big deal, a guy holding things in front of a projector... That's what I thought at first. But what I can't figure out is why doesn't the image show on the guys hands and face. You'd think at least some light would be reflected off of his hands.

Re:Still can't figure out how it works...

[lommer]

Ya, I'm confused about how it works too. My only thought was that maybe the projector projects an image using IR and the device flouresces with the appropriate colour. That way the projection would be invisible on him.

I have no idea whether this is how they actually do it, it's just a thought.

Re:Still can't figure out how it works...

[uradu]

> why doesn't the image show on the guys hands and face.

What guy?

Re:Still can't figure out how it works...

[Orne]

When you watch football, and you see the yellow line drawn on the screen, do you wonder why it looks like it goes under their feet and not through their bodies? Think of it the same way the local weatherman stands in front of the map... real-time color analysis/replacement.

1. Take a picture of the background.
2. Put an object in front of the camera, painted with an obviously different color than the objects in the background.
3. Profit... no wait, I mean... using a computer, replace all the pixels of that special color with the pixels of the background image, and you have modern television overlays.
4. But this only works through a video feed; to move into real-space, you need a video projector. Take the image with the special color, and use it as a mask... the image will only shine where the special color exists, and not shine where it doesn't. That's how they get the projector to not shine on the hands.
5. Note that their special color is silver; like a white screen, a sufficiently light color such that color projected on it will appear. Viola!

Re:Still can't figure out how it works...

[Jonny 290]

Note that their special color is silver; like a white screen, a sufficiently light color such that color projected on it will appear. Viola!


If you use a different color, could you get, say, a cello, or perhaps a double bass?

Re:Still can't figure out how it works...

[alphaseven]

Answering my own question here (guess I'm the first person to read the pdf).

We used a pinhole as the projector's iris in order to obtain a perfectly focused image. Furthermore, the projected image through the small aperture on the normal surface is too dim to be perceived by human eyes.

However the light coming out from the projector is reflected on the half mirror then on the screen and goes straight back in the eye to form the image, which is about ten or hundred times brighter than the image on the normal surface. Therefore the image only appears on the retroreflective material so that the viewer can observe as if the images projected on the retroreflective material are occluded by the object which exists in front of the screen.

So the image projected is too dim to be seen, but the objects are covered in a special highly reflective material (not just colored lightly like a projector screen).

I guess it might be sort of like a dim flashlight hitting a bicycle reflector at night, you only see the reflector lit up.

Re:Still can't figure out how it works...

[AJWM]

Nope, nope, nope. You give a fair description of the chromakey process (aka blue-screen or green-screen), but that won't work with a non-color like silver.

This is a variation on what's known as front-projection, invented back in the 1950s (by SF author William F. Jenkins, better known by his pen-name Murray Leinster) and used in "2001: A Space Odyssey" to film the "Dawn of Man" sequences.

The trick is the retro-reflective material (Scotchlite, originally). You can either use that as a backdrop or, as in this case, you coat an object or cloak with it. It's made up of tiny beads that reflect light back in the direction it came from. (The same stuff is used on the ping-pong balls that attach to your joints in a motion-capture setup -- the video camera doing the mo-cap is co-located with a light source of a particular color so that the balls show up as bright spots in that color).

Anyway, if you project an image using low-intensity light, that image will be washed out by ambient light to most observers. To somebody (or a camera) next to the projector, that image will be quite bright when bounced off the retroreflective material. In "2001", the scenes of the ape-men with the African landscape in the background were all filmed on a sound stage, with the landscape projected onto a front-projection screen behind them.

To do the invisibity trick, you set up a piece of the screen in the foreground -- whatever moves behind it will seem to have gone invisible.

(The half-mirror in the diagram is just there so that the camera and projector don't have to be physically in the same spot.)

That is so cool!

[tswinzig]

Now we just need to get the enemy to aim these giant projectors at us!

Re:holy sh*t

[lommer]

Ya, and it also requires that you have a half-mirror (one of those ones the cops use to watch interrogation rooms) in between you and the viewer. That is fine when you are dealing with a camera, but it isn't practically applicable to most areas that this would be useful in.

Re:Theator applications

[lommer]

I dunno, the half-way mirror thing kind of limits this application. Not to mention the parralax problems you'll get with an audience that is spread out.

Slashdot requires you to wait 2 minutes between each successful posting of a comment to allow everyone a fair chance at posting a comment.

A first in Slashdot history...

[Romothecus]

Not only have the editors forgotten that this has already been on Slashdot, apparently, the wonderful community hasn't noticed yet either.

That's because ITS FROM GHOST IN THE SHELL

[Romothecus]

Look at the page again. Scroll to the bottom:

Publications

M. Shiro, Ghost in the Shell, Kodansya, 1991

M. Inami, N. Kawakami, D. Sekiguchi, Y. Yanagida, T. Maeda and S. Tachi, Visuo-Haptic Display Using Head-Mounted Projector, Proceedings of IEEE Virtual Reality 2000, pp.233-240, 2000(vr2000.pdf(350k))

M. Inami, N. Kawakami, Y. Yanagida, T. Maeda and S. Tachi, Method and Device for Providing Information, US PAT. 6,341,869 , 2002

I'm not sure if that means that Ghost In the Shell was the inspiration or an actual source or what... but clearly the connection is not accidental.

gg, try to look at the whole page next time.

This is maybe halfway there

[cacav]

I will admit that this works better than I expected from the video clips. But I think it's barely halfway towards the goal of personal optical camouflage. From what they describe, it requires an external display source, and applies a bluescreen style approach.

What they need to do is figure out a way to create a piece of clothing which *is* the display, and not merely a reflector or a bluescreen like what they have now. Maybe take this foldable display technology I've read about (mainly geared towards disposable displays), and attach it to fabric. Then make a uniform out of it. Then, you can determine the image behind the person and display it on the uniform.

They're already halfway there, since they can determine what's behind the person and display it on a special colored material. Now all they need is a fabric that can display images, and then they can transmit the background onto the uniform.

Then we're talking about optical camouflage.

"The results are probably better than you would.."

[Superfreaker]

I expect it to work as well as it did in James Bond.

Instead of using projectors though, Q explained that the car was made up of a material that acted like a screen. I guess like a plasma screen, or even that new electronic paper, it may not be that far off.

It would be much better than a projector...hmmm...

Stereoscopy, too

[cryptor3]

Quoting the english version of the page,

Moreover, to project a stereoscopic image, the observer looks at the masking object more transparent.

While the grammar is somewhat ambiguous, I believe that what they're saying is that they attempt to project a stereoscopic image of the background onto the observer's eyes. However, it brings up the possibility of using a two-camera analysis of the scene to figure out where the silver coat is.

I love the references they made

[mselmeci]

* M. Shiro, Ghost in the Shell, Kodansya, 1991

I love it how they credited manga for the idea.

ohhhhh-kaaaay

[Twirlip of the Mists]

Two giant problems jump right out at me.

One: depth perception. Because we're watching movies of the demonstration instead of seeing it in person, it looks fairly good. But if you saw this technique in person, with binocular vision, you would be able to tell without any difficulty what you were seeing. If the background is sufficiently cluttered-- like a jungle backdrop-- and if the camouflaged thing is stationary, it might be good enough to fool a passing glance. But then again, so is your average ghilie suit.

Two: parallax. (This has been mentioned upthread already.) The demonstrations are all shot with a stationary camera. If the camera were to move, the illusion would be ruined. I'd imagine from looking at the illustrations that the geometry of the projector, reflector, mirror, and observer are all pretty critical. If the observer takes a step in any direction, the illusion will be broken.

All in all, I can't see how this could ever be practical. The most important flaw is that it's not even remotely adaptive. You have to have a perfect still photograph of your background to project against the mirror. If a bird flies behind the camouflaged object or person, the game is up.

Neat trick, but not very impressive.

Why not a simpler method?

[Simon Field]

Since this scheme is only good for hiding from one watcher at a time, and only from a one-eyed watcher when close, it seems that a simpler method will work.

Hold a rear-projection screen in front of you as a shield. It will have a pole coming toward you, that is long enough so the end of it is behind you when you hold the shield. On this end is a camera looking at the scene behind you.

Now it works without half-silvered mirrors, it is portable (you carry it around), and it works even if the background is moving.

Re:Why not a simpler method?

[Kaz Riprock]

...and then you get shot from the side and your enemy takes your camo and gun.

Re:Why not a simpler method?

[Simon Field]

We started by saying that the initial design only hid you from one direction. Both designs have the flaw you point out.

But if I were trying to hide a spy plane over Iraq, I might be able to assume that I am hiding from people on the ground, since we have already prevented Iraqi presence to my sides.

Re:"The results are probably better than you would

[quintessent]

Yeah, after reading the post, the results were worse than I expected. Oh well.

Re:"The results are probably better than you would

[Hast]

Well you see, the trick they used in the James Bond movies was that /there was actually no car there/. It's generally a lot easier add an invisible car than to remove a visible one.

Yeah yeah, I'm being mildly sarcastic here. My point being that if you actually made the "Vanish" IRL the result would be more similar to what this guy has than to the the digitally processed movie.

Re:We're now offtopic but...

[Qrlx]

yes it will. someone will mod this down.