It looks like you're using an Ad Blocker.

Please white-list or disable AboveTopSecret.com in your ad-blocking tool.

Thank you.

 

Some features of ATS will be disabled while you continue to use an ad-blocker.

 

Researcher's Analysis of al Qaeda Images Reveals Surprises

page: 1
9
<<   2 >>

log in

join
share:

posted on Aug, 2 2007 @ 07:18 AM
link   

Researcher's Analysis of al Qaeda Images Reveals Surprises


blog.wired.com

Neal Krawetz, a researcher and computer security consultant, gave an interesting presentation today at the BlackHat security conference in Las Vegas about analyzing digital photographs and video images for alterations and enhancements.

But more interesting were the examples Krawetz gave of al Qaeda images. Krawetz took an image from a 2006 al Qaeda video of Ayman al-Zawahiri (above right), a senior member of the terrorist organization.
(visit the link for the full news article)




posted on Aug, 2 2007 @ 07:18 AM
link   
Very interesting. Keep in mind that Blackhat conferences attract some of the finest minds in the IT security world. The creator of this application does note that there are many possibilities as to why several images of Ayman al-Zawahiri & Azzam al-Amriki may have been doctored. Could the images be part of a coding mechanism as stated in the article or are we seeing poor workmanship from digitial editing by our friends at either the CIA or NSA. Could this be proof of a government influenced fear campaign?

blog.wired.com
(visit the link for the full news article)

[edit on 2-8-2007 by brill]



posted on Aug, 2 2007 @ 08:21 AM
link   
Is there any release info on this software? Name or Date? Its about time something like this became commercially available. Maybe this software can put some of the 9/11 debate to rest?



posted on Aug, 2 2007 @ 08:29 AM
link   
None that I can find so far. It was probably an exclusive release to Blackhat members but may be released to the general public soon.

brill



posted on Aug, 2 2007 @ 09:29 AM
link   
interesting read. Thanks for posting it.

I think the videos in question were altered by the CIA, and I dont trust the translations, but thats just me.

Believe half of what you see and none of what you hear.



posted on Aug, 2 2007 @ 09:30 AM
link   
Sounds pretty interesting, I'd be interested to hear more about this program.

As for the first picture shown, Al-Zawarhiri (?), you don't need a program to tell the the background was added in. Firstly it just looks fake but if you look at the sides of his beard you'll see that they're blue. Blue is one of the colors used in a technique called "Chroma-key". Although Hollywood uses neon green since it's such an odd artificial color that the end result is cleaner.

All you have to do is film all the action in front of a blue/green screen on one camera and have the background on a different feed. Same thing the weather man uses.

I still don't really understand how the program works or why it returns a visual result but to me it seems the fake elements in the photos could be picked out with a well trained eye.

You'd also need the original photo to determine how it was faked since the program only gives you info on the last program or tool used on it. If you copy the picture from a website it's surely been edited somehow.



posted on Aug, 2 2007 @ 11:29 AM
link   
Two main possibilities on the possible manipulation in question:

1. Al-Qaeda is "spiffing up" their videos for purely presentational reasons.
2. NSA/CIA are involved in fakery.



posted on Aug, 2 2007 @ 11:31 AM
link   
The article was updated and the source code is available on the site now.

Link to source code

Does anyone know how to compile it to give it a try?



posted on Aug, 2 2007 @ 11:39 AM
link   
Steganography software is available, a link to a topology.org page... scroll down to the "steganography section".

I've not done much if any steg research but prowling around Outguess StegHide and Hydan... have supporters. There are some info links as well on the topology.org page here.

Cheers,

Vic



posted on Aug, 2 2007 @ 11:50 AM
link   
Someone take a shot at those rajman, chad and ty drone images with that software?



posted on Aug, 2 2007 @ 11:56 AM
link   

Originally posted by uberarcanist
Two main possibilities on the possible manipulation in question:

1. Al-Qaeda is "spiffing up" their videos for purely presentational reasons.
2. NSA/CIA are involved in fakery.


I guess that about sums it up really.

Nice find and great software. I suggest the government hire the creator and keep the software for their own personal use. Can't have civilians finding discrepancies in officially documented photos.



posted on Aug, 2 2007 @ 12:09 PM
link   

Originally posted by PsykoOps
Someone take a shot at those rajman, chad and ty drone images with that software?


Exactly! Now if I only knew how to compile the software so I can make it run. Does anyone know how to do this, or point me in the right direction?

Thanks.



posted on Aug, 2 2007 @ 12:16 PM
link   
New information is great and all but in the end this whole thing doesn't tell us squat. Yep, that background is fake - it has all the hallmarks of a bad computer generated virtual set. Why you would need software to alert you something that obvious is beyond me.

If fact, before even reading the article I saw that photo and said to myself "Ugh, throw some grain on that background and maybe add a little texture and lighting variation to the lampshade."

What is more interesting to me is how the researcher avoided examining any photography that could be in the least bit interesting. I mean really, "Oh no! They are adding books to their videos? Those terrorists are an even bigger threat to freedom now."

I'm going to have to download the code though. I can see making this an automated system on a server somewhere (maybe ATS) and having people submit photos for real-time analysis. Would be interesting to see what portion of all those 911 and JFK photography or even Google's maps are faked or edited.

Jon



posted on Aug, 2 2007 @ 12:19 PM
link   
i'll try to get the source and compile it to see what it does. i'll keep everyone posted.



posted on Aug, 2 2007 @ 12:24 PM
link   
That idea for web based client is awesome, I think ATS should sponsor that definately. I dont know how much resources that would take, but since ATS is upgrading it's servers anyway... Springer!!!



posted on Aug, 2 2007 @ 12:27 PM
link   
It just makes sense to me that if A-Q or Talibans' or Jihadist groups

commenced to make video footage that featured any of the groups
leadership...they themselves would screen the footage and deliberately
enhance/distort/alter anything that might be analysed for intelligence data.

a new mole on a face, the length/color of face hair which may date the
video image, etc etc etc.
the remaster of the video images alone tells the U.S. intel agencies
something of the aggressors capabilities and prowess,
and a next generation capability would itself help date the ?new?
video release in relation to actual/real world dates.

i think all this supposing the CIA or black-ops is actually behind
the "terrorists" warnings & fatwas is ....well....Arrogant on our part- -
by creating the aura that only 'we' are capable of such trickery &
that the rest of the world & the avowed enemy cannot be so capable.



posted on Aug, 2 2007 @ 12:29 PM
link   
this is the result i got from one of the chad drone images:

#File: ./chad/2-2.jpg

Quantization table
Precision=0; Table index=0 (luminance)
3 2 2 3 2 2 3 3
2 3 3 3 3 3 4 7
5 4 4 4 4 9 6 7
5 7 10 9 11 11 10 9
10 10 12 13 17 14 12 12
16 12 10 10 14 20 15 16
17 18 19 19 19 11 14 20
22 20 18 22 17 18 19 18
Estimated quality level = 89.46%

Quantization table
Precision=0; Table index=1 (chrominance)
3 3 3 4 4 4 8 5
5 8 18 12 10 12 18 18
18 18 18 18 18 18 18 18
18 18 18 18 18 18 18 18
18 18 18 18 18 18 18 18
18 18 18 18 18 18 18 18
18 18 18 18 18 18 18 18
18 18 18 18 18 18 18 18
Estimated quality level = 84.19%
Average quality: 91.11% (91%)


doesn't really tell us much. i think this is just analyzing the amount of compression for the image. doesn't provide any big clues into the validity of a photo.

[edit on 2-8-2007 by an0maly33]



posted on Aug, 2 2007 @ 01:07 PM
link   
It only shows one 8x8 q-matrix? That is odd - each of those should be an 8x8 region of pixels. I haven't compiled it yet as I am waiting on the download to finish (didn't notice that I never installed a compiler on this box, hah!)

I would expect there to be a couple hundreds of little 8x8 regions in even the smallest image. Then you have to look for differences in the estimated quality level between multiple 8x8 regions. At least that is my understanding of how this thing works.

Jon



posted on Aug, 2 2007 @ 01:14 PM
link   
i'll have to mess with known real images vs altered ones to see how the results differ. just to clarify what this program is doing: it seems like it tries to check the variation in compression throughout the image. for example if i have a jpeg, paste a ufo on to it and resave it, the original data has now been compressed twice and the ufo was only compressed once. that discrepency tells you that the image is fake. but i fail to see how the data we're getting from this program's output tells us anything useful at this point.

one thing i am noticing as i mess with this program: pictures taken from the same camera have the same variances. the results on 3 pictures i analyzed are identical. the chad drone images from this thread:

www.abovetopsecret.com...

do not all have identical signatures - 2 of them do.

[edit for more content]


[edit on 2-8-2007 by an0maly33]



posted on Aug, 2 2007 @ 01:44 PM
link   
After trying it out myself I can reveal that this code does nothing but print out the two or three quantization tables.

How does that tell you anything?

Well it doesn't really, the only thing that can tell you is what was the last program or camera to compress the image. Those tables are hard-coded into software or hardware allowing you to use it like a fingerprint.

Basically, you take a photograph and extract the q-tables then compare them with some of the tables listed here:
Index of Quantization Table Sources

If you find any that match you can be pretty sure that was the source. I am willing to bet 96% of the images found on the internet have Photoshop tables. Which means nothing in the long run.

That same website provides an easy to use windows utility called "JPEGSnoop" to help you extract the tables from an image: JPEGSnoop

NOTE: The code provided is not at all what they show in the article from the OP. That is a combination of techniques (mostly principle component analysis) to show differences in a photo which, while interesting, was not released to the public by the presenter.

Jon




top topics



 
9
<<   2 >>

log in

join