Detection of Copy/Move Forgery in Images

first of all, i'm not quite sure where to put this so i'm putting this here coz i feel this is the part of ats where most people who are inclined to computers would most likely congregate. i found this link through reddit and i feel that it is interesting enough to bring this to your attention just in case one of you guys and gals want to give this a try.

Digital images are easy to manipulate and edit due to availability of powerful image processing and editing software. Nowadays, it is possible to add or remove important features from an image without leaving any obvious traces of tampering. As digital cameras and video cameras replace their analog counterparts, the need for authenticating digital images, validating their content, and detecting forgeries will only increase. Detection of malicious manipulation with digital images (digital forgeries) is the topic of this paper. In particular, we focus on detection of a special type of digital forgery – the copy-move attack in which a part of the image is copied and pasted somewhere else in the image with the intent to cover an important image feature. In this paper, we investigate the problem of detecting the copy-move forgery and describe an efficient and reliable detection method. The method may successfully detect the forged part even when the copied area is enhanced/retouched to merge it with the background and when the forged image is saved in a lossy format, such as JPEG. The performance of the proposed method is demonstrated on several forged images.

Link (warning: pdf)

Implementation

so you think one of you can create a point and click app for the rest of us so that we can analyze all those pics that purportedly have gremlins, aliens and ufos in them?

[edit on 7.1.10 by toreishi]

In one of my media design courses, we used a lot of greyscale shifts, cloning tools and layer tools to manipulate images. Its really amazing how easy it is to falsify digital imagery today, even with the free and share software.

However, I see the problem when you are trying to reverse the process. The algorithms, as I understand them, look for exact similarities at the pixel level to match exact pixel shifts or copies to detect where the lie in the image might be. Depending on the level of depth of the image, it may be quite possible to find, but I doubt very easy to undo, depending on how the artist developed the image.

If layers are compressed repeatedly into one another, then they are bonded and whatever they are hiding could very likely be scrubbed away with the overlay. While it would tell you there is a fake image there, it may be more difficult, or even impossible, to undo it and see what is hiding.

Its a love hate relationship with technology of late.

there are a few things you can do without going through much trouble

One is to check the exif data of an image, a true unedited photograph will have everything about the camera when it was taken on(date is very useful in debunking) shutter speed etc.

But most importantly it will tell you if it has been edited in photoshop or other programs

You can get a free addon on firefox that will give you the exif data on any pic online with just a click.

Another usefull tool, is Tineye, the reverse image search. It will find an exact match of the image, and show you where else it has been posted online.

[edit on 2-7-2010 by MR BOB]

Originally posted by MR BOB
One is to check the exif data of an image, a true unedited photograph will have everything about the camera when it was taken on(date is very useful in debunking) shutter speed etc.

Online EXIF viwer - All though it can't be too hard to edit or simply remove this data - still useful.