Applied Visual Cryptography

[+Osmirik]

I've been wondering about a scheme for a multi-cache that would allow people to find any two of three first-stage caches, and only then be able to determine the location of the final cache.

There's a system called "Visual Cryptography", which allows you to create some seemingly random images, that, when printed onto transparencies and lined up, will produce a recognisable image. I'm thinking of creating a set of three of these images, such that any two will show the co-ordinates of the final cache when combined, then printing lots of copies onto transparencies, cutting them up, and putting them in microcaches three seperate locations. (Eventually - this probably won't be my first cache!)

I've got an example you can try in your browser here - just drag any image over any other image. Make sure to line them up exactly to see the final image.

 

Thoughts? Impressions? Would this make a cool cache? Am I likely to run into issues maintaining a stock of bits-of-transperency? Any other issues with this idea?

Feedback would be appreciated.

[+SixTen]

Sounds like a good idea...instead of using the actual transparencies, you could upload the images to a site and give the web location at the cache...not sure how well that would work out with what you had in mind...but it would save on the transparencies...

 

-Six

[+Osmirik]

Sounds like a good idea...instead of using the actual transparencies, you could upload the images to a site and give the web location at the cache...not sure how well that would work out with what you had in mind...but it would save on the transparencies...

 

-Six

True, but then why bother with the visual crypto at all?

[+SixTen]

True

[+bigredmed]

How would you have micro 1's transparency available at micro 2? Would you make multiple copies? If so, how would you monitor for people grabbing all the copies of a given transparency?

[+Divine]

Thoughts? Impressions? Would this make a cool cache? Am I likely to run into issues maintaining a stock of bits-of-transperency?

The idea generally is cool, but like you suggested, you may run into maintaining issues, depending how geocacher-dense area you live in. Better have the caches close to your house or other place you frequent.

 

Another idea is to print only few transparencies, and obligate the cachers to bring them back to where they got them. This should be made clear in the cache description already.

 

You could hide the first cache near the obvious parking spot. Place few transparencies here with an accompanying instructions to take one transparency with and go find the next caches either Here (set of coords) or There (another set of coords). The point is, you give the next caches' coords not earlier than at this first container, so the searchers don't know them beforehand.

 

Here and There would have hidden containers with different cryptographic images, printed on paper. When the cachers find either of them, they just place the transparency they took from the first container on the paper, and they get the final coords for the cache.

 

When they've found the cache and are coming back to their car/whatever, they're supposed to drop by at the first container and leave the transparency there. You can place more than one transparency there, should more than one geocachers happen to come there at the same time.

 

I know this wasn't quite the same idea you originally had in mind, but this could save you from some maintenance visits to the caches. Feel free to elaborate.

[+greengolftee87]

you could also ask them to leave them in the final cache where you could go and collect them

[+Robespierre]

This seems really great - the possibilities. Now, I usually shy away from the "too involved" caches, like 6-step multis, but I can see a two-day cache here, where you need time to set it down and figure it out. worth it 'cause it is very cool!

[Fakk 2]

I like this, make it so that if the transparencies aren't overlapped properly. eg: the top is not set up on the top of the other (upside down) you get a different coordinate. This would make it more of a challenge.

[+texastracker]

I really like this idea! I do agree that the maintenance could be overwhelming but the cacher's I've met would gladly take care of the cache for the next guy.

[CoyoteRed]

Sounds like a maintenance nightmare.

 

First you'd have to make sure each of the transparencies are marked so the A series sheet don't get mixed with the B series cache. Otherwise, you'd only need to find that one cache with multiple versions of transparencies to go to the final cache.

 

Then then there's the keeping it restocked. Either you'd have to do it yourself or rely on others to do it for you.

 

What might be a little less of a maintenance chore is have only one cache have these transparencies. Each sheet will have several images and a set of coordinates. Go to the next location and align the puzzle with another sheet that stays in the cache (printed on regular paper). The cacher follows the clues that way and only has to keep track on one sheet. You'd only have to worry about one cache being stocked with transparencies and there's little chance of them getting mixed up in other caches.

 

In fact, with this technique you could print the intermediate images on a weatherproof adhesive-backed paper and stick them in various places--behind signs, lampposts, etc. No need for actual containers. I had an idea for those 2d barcodes but couldn't figure out a way for others to read it without expensive equipement. This is kind of an extension.

 

Good luck on your idea!

[+Renegade Knight]

Cool concept. I like it. You will need to flat out plan on stocking your micro's with clues every now and again. Cachers have to take them to solve the puzzle and even if they all put them back whey they are done there will be a certain number that are always in circulation and not in the caches as clues.

[+Stunod]

You could have the cacher print out one image at home and match it up with a transparancy that stays in the cache. I'm not sure if there would be problems with different printers resulting in mismached sizes of images.

[+Seamus]

You could have the cacher print out one image at home and match it up with a transparancy that stays in the cache. I'm not sure if there would be problems with different printers resulting in mismached sizes of images.

Include a scale bar in the image for the user to print out at home. They may need to print it a couple of times before the bar is 2 inches long (or whatever), but I would think that it should work fairly well that way.

[+Renegade Knight]

You could have the cacher print out one image at home and match it up with a transparancy that stays in the cache.  I'm not sure if there would be problems with different printers resulting in mismached sizes of images.

Include a scale bar in the image for the user to print out at home. They may need to print it a couple of times before the bar is 2 inches long (or whatever), but I would think that it should work fairly well that way.

dadgum you guys are good. That's something I should of thought of.

[+PandyBat]

It sounds like a great idea if you can get it worked out right. I'd like to do one like this too but I'm not that great with computer graphics. Awesome idea for a cache, Osmirik!

[+rusty_tlc]

You could have the cacher print out one image at home and match it up with a transparancy that stays in the cache.  I'm not sure if there would be problems with different printers resulting in mismached sizes of images.

Include a scale bar in the image for the user to print out at home. They may need to print it a couple of times before the bar is 2 inches long (or whatever), but I would think that it should work fairly well that way.

dadgum you guys are good. That's something I should of thought of.

This is the same method I use to verify electronic files. You might also add feducials for alignment.

[+Seamus]

Maybe a larger grain size, too. I just tried it with printouts made from your online example, and the alignment was pretty tricky, even using a flat surface. Doing it out "in the field" would have been far more difficult.

 

Larger grains and a couple of registration targets though, and I think you're well on your way to having an outstanding Geocache! Let me know when you get it nailed down and placed - I'd be interested in seeing how people do with it.

[+Tobias Maximus]

What an awesome idea! How did you make those images?

[+wildearth2001]

Thoughts? Impressions? Would this make a cool cache? Am I likely to run into issues maintaining a stock of bits-of-transperency?

The idea generally is cool, but like you suggested, you may run into maintaining issues, depending how geocacher-dense area you live in. Better have the caches close to your house or other place you frequent.

 

Another idea is to print only few transparencies, and obligate the cachers to bring them back to where they got them. This should be made clear in the cache description already.

 

You could hide the first cache near the obvious parking spot. Place few transparencies here with an accompanying instructions to take one transparency with and go find the next caches either Here (set of coords) or There (another set of coords). The point is, you give the next caches' coords not earlier than at this first container, so the searchers don't know them beforehand.

 

Here and There would have hidden containers with different cryptographic images, printed on paper. When the cachers find either of them, they just place the transparency they took from the first container on the paper, and they get the final coords for the cache.

 

When they've found the cache and are coming back to their car/whatever, they're supposed to drop by at the first container and leave the transparency there. You can place more than one transparency there, should more than one geocachers happen to come there at the same time.

 

I know this wasn't quite the same idea you originally had in mind, but this could save you from some maintenance visits to the caches. Feel free to elaborate.

now thats a great idea

[+Osmirik]

A few ideas I hadn't considered here - I'd considered providing one cache with pieces of paper for people to take, and providing a transparency at other caches for them to match it with, so I'd only have to stock paper, rather than the reverse. A system could also be devised so the same piece of paper decodes to different images depending on what micro you go to.

I hadn't considered printing the intermediate stages on waterproof stickers and sticking them to things - brilliant idea! They'd probably even be mistaken for some sort of barcode for geomuggles .

I also hadn't considered getting people to print the parts themselves, but you'd have to get them exactly right - and even then I'd have to use larger pixels. As it stands, even the tiniest misalignment will render parts of the images unreadable.

And then there's the suggestion of making them show a different image if they align differently - I'll have to look into that, but I suspect it's unfortunately impractical.

Toby: The technique used is an extension of the one described here. I had to work out how to do a 2 in 3 system myself, as the scientific papers on all the extensions (embedding the individual images in pictures, improving contrast, multiple keyshares, etc etc) all seem to involve ridiculously complex equations - beyond my current math understanding. :/

 

Edit: Is sticking waterproof stickers to things 'defacing public property' if they're not anywhere obvious?

Edited June 17, 2004 by Osmirik

[Dinoprophet]

I have a cache with photos in each stage. Maintenance has been a pain, but only because stages keep disappearing. Keeping up the pictures themselves hasn't been a problem. I have an envelope in the final cache for people to leave the pictures. If it was hit more than 10 times in a single wekend, there'd be a problem, but that's only a very remote possibility.

 

What I'm saying is, I don't think maintenance will be a problem. And I like the idea.

[+Corp Of Discovery]

This idea sounds good, what about taking it low tech tho. Could you use graph paper and color the individual sheets? This would of course be easier with an image that is not so complex (could use smaller squares for more complex if needed) but should work with numbers easily enough, and being the individual squares are bigger than your example, aligning them would be easier too. With 2 sheets you should be able align them easily (put a note saying top on top of sheets) and be able to hold them to a light source and see the intended pattern/message/coords.

[+Osmirik]

Could you use graph paper and color the individual sheets?

I'm sorry, I don't get what you're getting at.

Also, a certain number of pixels width and height of the lines used to form characters is needed to make them easy to distinguish from the background noise.

[+Corp Of Discovery]

Could you use graph paper and color the individual sheets?

I'm sorry, I don't get what you're getting at.

Also, a certain number of pixels width and height of the lines used to form characters is needed to make them easy to distinguish from the background noise.

Think of each pixel as a square on the graph paper. Don't use the background noise tho, that will keep it simpler.

[+Osmirik]

Think of each pixel as a square on the graph paper. Don't use the background noise tho, that will keep it simpler.

Unfortunately, it doesn't work like that - without the 'background noise', the image would be clearly visible on each part without having to combine them.

[+Corp Of Discovery]

Think of each pixel as a square on the graph paper. Don't use the background noise tho, that will keep it simpler.

Unfortunately, it doesn't work like that - without the 'background noise', the image would be clearly visible on each part without having to combine them.

But what I am proposing would be to break the image itself into two or more parts that when assembled would show the complete image.

[+Osmirik]

Think of each pixel as a square on the graph paper. Don't use the background noise tho, that will keep it simpler.

Unfortunately, it doesn't work like that - without the 'background noise', the image would be clearly visible on each part without having to combine them.

But what I am proposing would be to break the image itself into two or more parts that when assembled would show the complete image.

But can you propose a 2 in 3 (2 out of 3 parts required to reconstruct the original) system with that?

[+Corp Of Discovery]

Think of each pixel as a square on the graph paper. Don't use the background noise tho, that will keep it simpler.

Unfortunately, it doesn't work like that - without the 'background noise', the image would be clearly visible on each part without having to combine them.

But what I am proposing would be to break the image itself into two or more parts that when assembled would show the complete image.

But can you propose a 2 in 3 (2 out of 3 parts required to reconstruct the original) system with that?

It should be able to be done, but I think it may be a bit more than I could chew. But why retrict yourself to 2 of 3? Make it so all 3 need to be found and when all 3 are put together you get the full image, or alternatively do it with only 2 caches and two parts needed. You may also want to list each cache individually- that way cachers could get credit for each stage of the multi. Those seem to be a bit more popular than the 4 parts one find kind (at least where I'm from)

[+BillP3rd]

I think it's a really cool idea! (I reviewed your reference to the technique and it's deceptively simple to implement!) I've seen multi-point caches where something was to be removed from one or more waypoints and taken along on the hunt. Obviously, it's up to the cache owner to maintain an adequate stock in the appropriate caches. Beyond that, go for it!!

 

BTW: Did you write the software to generate your example yourself or use something from the 'net? I've considered doing something similar with a "Magic Eye" type 3-dimensional image of the sort that were popular a few years back.

Edited June 18, 2004 by BillP3rd

[+Osmirik]

I think it's a really cool idea! (I reviewed your reference to the technique and it's deceptively simple to implement!) I've seen multi-point caches where something was to be removed from one or more waypoints and taken along on the hunt. Obviously, it's up to the cache owner to maintain an adequate stock in the appropriate caches. Beyond that, go for it!!

 

BTW: Did you write the software to generate your example yourself or use something from the 'net? I've considered doing something similar with a "Magic Eye" type 3-dimensional image of the sort that were popular a few years back.

Yes, I wrote the software - a simple PHP script. It's somewhat buggy, but feel free to have a play at http://www.notdot.net/viscrypt/viscrypt2.php. A 2 of 2 system (the same one originally described in that link) is available at http://www.notdot.net/viscrypt/viscrypt.php.

 

It should be able to be done, but I think it may be a bit more than I could chew. But why retrict yourself to 2 of 3? Make it so all 3 need to be found and when all 3 are put together you get the full image, or alternatively do it with only 2 caches and two parts needed. You may also want to list each cache individually- that way cachers could get credit for each stage of the multi. Those seem to be a bit more popular than the 4 parts one find kind (at least where I'm from)

 

One of the reasons I originally wanted to implement the system, besides the 'cool' factor was that someone wouldn't have to find all the caches, only a subset. If they found one particularaly hard, or one was missing, it'd still be finishable.