Free Republic
Browse · Search
General/Chat
Topics · Post Article

Skip to comments.

Digital Forensics: 5 Ways to Spot a Fake Photo
Scientific America ^ | June 2, 2008 | Hany Farid

Posted on 06/04/2008 5:39:02 PM PDT by yankeedame

Features - June 2, 2008

Digital Forensics: 5 Ways to Spot a Fake Photo

By Hany Farid

This story is a supplement to the feature "Digital Forensics: How Experts Uncover Doctored Images" which was printed in the June 2008 issue of Scientific American.

Lighting

Composite images made of pieces from different photographs can display subtle differences in the lighting conditions under which each person or object was originally photographed. Such discrepancies will often go unnoticed by the naked eye.

For an image such as the one at the right, my group can estimate the direction of the light source for each person or object (arrows). Our method relies on the simple fact that the amount of light striking a surface depends on the relative orientation of the surface to the light source. A sphere, for example, is lit the most on the side facing the light and the least on the opposite side, with gradations of shading across its surface according to the angle between the surface and the direction to the light at each point.

To infer the light-source direction, you must know the local orientation of the surface. At most places on an object in an image, it is difficult to determine the orientation. The one exception is along a surface contour, where the orientation is perpendicular to the contour (red arrows right). By measuring the brightness and orientation along several points on a contour, our algorithm estimates the light-source direction.

For the image above, the light-source direction for the police does not match that for the ducks (arrows). We would have to analyze other items to be sure it was the ducks that were added.

Eyes and Positions Because eyes have very consistent shapes, they can be useful for assessing whether a photograph has been altered.


A person’s irises are circular in reality but will appear increasingly elliptical as the eyes turn to the side or up or down (a). One can approximate how eyes will look in a photograph by tracing rays of light running from them to a point called the camera center (b). The picture forms where the rays cross the image plane (blue). The principal point of the camera—the intersection of the image plane and the ray along which the camera is pointed—will be near the photograph’s center.

My group uses the shape of a person’s two irises in the photograph to infer how his or her eyes are oriented relative to the camera and thus where the camera’s principal point is located (c). A principal point far from the center or people having inconsistent principal points is evidence of tampering (d). The algorithm also works with other objects if their shapes are known, as with two wheels on a car.

The technique is limited, however, because the analysis relies on accurately measuring the slightly different shapes of a person’s two irises. My collaborators and I have found we can reliably estimate large camera differences, such as when a person is moved from one side of the image to the middle. It is harder to tell if the person was moved much less than that. 

Specular Highlights

Surrounding lights reflect in eyes to form small white dots called specular highlights. The shape, color and location of these highlights tell us quite a bit about the lighting.

In 2006 a photo editor contacted me about a picture of American Idol stars that was scheduled for publication in his magazine (above). The specular highlights were quite different (insets).

The highlight position indicates where the light source is located (above left). As the direction to the light source (yellow arrow) moves from left to right, so do the specular highlights.

The highlights in the American Idol picture are so inconsistent that visual inspection is enough to infer the photograph has been doctored. Many cases, however, require a mathematical analysis. To determine light position precisely requires taking into account the shape of the eye and the relative orientation between the eye, camera and light. The orientation matters because eyes are not perfect spheres: the clear covering of the iris, or cornea, protrudes, which we model in software as a sphere whose center is offset from the center of the whites of the eye, or sclera (above right).

Our algorithm calculates the orientation of a person’s eyes from the shape of the irises in the image. With this information and the position of the specular highlights, the program estimates the direction to the light. The image of the American Idol cast (above;

directions depicted by red dots on green spheres) was very likely composed from at least three photographs. 

Send in the Clones Cloning—the copying and pasting of a region of an image—is a very common and powerful form of manipulation.

This image is taken from a television ad used by George W. Bush’s reelection campaign late in 2004. Finding cloned regions by a brute-force computer search, pixel by pixel, of all possible duplicated regions is impractical because they could be of any shape and located anywhere in the image. The number of comparisons to be made is astronomical, and innumerable tiny regions will be identical just by chance (“false positives”). My group has developed a more efficient technique that works with small blocks of pixels, typically about a six-by-six-pixel square (inset).

For every six-by-six block of pixels in the image, the algorithm computes a quantity that characterizes the colors of the 36 pixels in the block. It then uses that quantity to order all the blocks in a sequence that has identical and very similar blocks close together. Finally, the program looks for the identical blocks and tries to “grow” larger identical regions from them block by block. By dealing in blocks, the algorithm greatly reduces the number of false positives that must be examined and discarded.

When the algorithm is applied to the image from the political ad, it detects three identical regions (red, blue and green).

Camera Fingerprints Digital retouching rarely leaves behind a visual trace. Because retouching can take many forms, I wanted to develop an algorithm that would detect any modification of an image. The technique my group came up with depends on a feature of how virtually all digital cameras work.

A camera’s digital sensors are laid out in a rectangular grid of pixels, but each pixel detects the intensity of light only in a band of wavelengths near one color, thanks to a color filter array (CFA) that sits on top of the digital sensor grid. The CFA used most often, the Bayer array, has red, green and blue filters arranged as shown below.

Each pixel in the raw data thus has only one color channel of the three required to specify a pixel of a standard digital image. The missing data are filled in—either by a processor in the camera itself or by software that interprets raw data from the camera—by interpolating from the nearby pixels, a procedure called demosaicing. The simplest approach is to take the average of neighboring values, but more sophisticated algorithms are also used to achieve better results. Whatever demosaicing algorithm is applied, the pixels in the final digital image will be correlated with their neighbors. If an image does not have the proper pixel correlations for the camera allegedly used to take the picture, the image has been retouched in some fashion.

My group’s algorithm looks for these periodic correlations in a digital image and can detect deviations from them. If the correlations are absent in a small region, most likely some spot changes have been made there. The correlations may be completely absent if image-wide changes were made, such as resizing or heavy JPEG compression. This technique can detect changes such as those made by Reuters to an image it released from a meeting of the United Nations Security Council in 2005 (above): the contrast of the notepad was adjusted to improve its readability.

A drawback of the technique is that it can be applied usefully only to an allegedly original digital image; a scan of a printout, for instance, would have new correlations imposed courtesy of the scanner.


TOPICS: Arts/Photography; Education; Miscellaneous
KEYWORDS: digitalimaging; fakebutaccurate; fakephotos; fauxtography; manufacturedmedia; photoforensics; photoshop; photoshopped
Not new, I know, but...what the hey?
1 posted on 06/04/2008 5:39:02 PM PDT by yankeedame
[ Post Reply | Private Reply | View Replies]

To: yankeedame

Bookmarked. Very good information. Thank you!


2 posted on 06/04/2008 5:52:59 PM PDT by EverOnward
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame

bump


3 posted on 06/04/2008 6:03:26 PM PDT by RaceBannon (Innocent until proven guilty; The Pendleton 8: We are not going down without a fight)
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame

4 posted on 06/04/2008 6:09:47 PM PDT by steveo (Time flies like an arrow, fruit flies like a banana.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame

Bump for later


5 posted on 06/04/2008 6:22:47 PM PDT by weegee (In 1988 Lenora Fulani was the 1st black woman to appear on presidential ballots in all 50 states)
[ Post Reply | Private Reply | To 1 | View Replies]

To: weegee

6 posted on 06/04/2008 6:51:07 PM PDT by BerryDingle (I know how to deal with communists, I still wear their scars on my back from Hollywood-Ronald Reagan)
[ Post Reply | Private Reply | To 5 | View Replies]

To: yankeedame
I have posted this many times and almost always people think its real.
7 posted on 06/04/2008 6:55:24 PM PDT by Eye of Unk (The world WILL be cleaner, safer and more productive without Islam.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame
Photobucket

This really happened.

8 posted on 06/04/2008 8:18:52 PM PDT by martin_fierro (< |:)~)
[ Post Reply | Private Reply | To 1 | View Replies]

To: martin_fierro

9 posted on 06/04/2008 8:21:34 PM PDT by JRios1968 ("If you go over a cliff with all flags flying, you are still going over a cliff"--Ronald Reagan)
[ Post Reply | Private Reply | To 8 | View Replies]

To: yankeedame
Hell, just look at Redbook magazine covers if you want an education in fakery.
10 posted on 06/04/2008 9:03:57 PM PDT by JoJo Gunn (Help control the girly-man population. Have the McCainiacs spayed or neutered.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: BerryDingle; steveo

11 posted on 06/04/2008 9:08:28 PM PDT by dead (I've got my eye out for Mullah Omar.)
[ Post Reply | Private Reply | To 6 | View Replies]

To: Eye of Unk
I have posted this many times and almost always people think its real.

Man, that commander better have his sailors clean the sand and dust off the deck!!!! JJJJ

12 posted on 06/05/2008 4:41:49 AM PDT by Arrowhead1952 (Typical white person, bitter, religious, gun owner, who will "Just say No to BO in Nov.")
[ Post Reply | Private Reply | To 7 | View Replies]

To: dead

13 posted on 06/05/2008 4:44:49 AM PDT by Arrowhead1952 (Typical white person, bitter, religious, gun owner, who will "Just say No to BO in Nov.")
[ Post Reply | Private Reply | To 11 | View Replies]

To: IncPen; Nailbiter

ping


14 posted on 06/05/2008 8:02:44 AM PDT by Nailbiter
[ Post Reply | Private Reply | To 1 | View Replies]

To: JoJo Gunn
Photobucket
15 posted on 06/05/2008 9:35:40 AM PDT by Sue Perkick (And I hope that what I've done here today doesn't force you to have a negative opinion of me....)
[ Post Reply | Private Reply | To 10 | View Replies]

To: yankeedame
Odd that they didn't use these examples:

Better to talk about ducks and presidential ads than to talk about the LA Times or Al-Reuters.

16 posted on 06/05/2008 9:42:29 AM PDT by weegee (In 1988 Lenora Fulani was the 1st black woman to appear on presidential ballots in all 50 states)
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame

BUMP


17 posted on 06/05/2008 9:42:35 AM PDT by Centurion2000 (Beware the fury of the man that cannot find hope or justice.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame

You can also get software from Canon and Nikon that produces a CRC for the image as it is first recorded in the camera. If, for some reason, you need to compare a published image with the original file produced in the camera, you can demonstrate that the original raw image is unaltered.


18 posted on 06/05/2008 9:43:18 AM PDT by js1138
[ Post Reply | Private Reply | To 1 | View Replies]

To: BerryDingle

He looks good in that one. He should buy it.


19 posted on 06/05/2008 9:44:13 AM PDT by weegee (In 1988 Lenora Fulani was the 1st black woman to appear on presidential ballots in all 50 states)
[ Post Reply | Private Reply | To 6 | View Replies]

To: yankeedame
HOW TO SPOT A FAKE PHOTO. Lesson one...


20 posted on 06/05/2008 9:47:00 AM PDT by weegee (In 1988 Lenora Fulani was the 1st black woman to appear on presidential ballots in all 50 states)
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame

21 posted on 06/05/2008 10:15:34 AM PDT by andy58-in-nh (Peace Is Not The Question.)
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame

22 posted on 06/05/2008 10:22:35 AM PDT by Lancey Howard
[ Post Reply | Private Reply | To 1 | View Replies]

To: yankeedame
Here is Part #1 of the classic

You Suck At Photoshop

If you have about a half-hour to kill, this series is priceless. Hilarious.

23 posted on 06/05/2008 10:27:23 AM PDT by Lancey Howard
[ Post Reply | Private Reply | To 1 | View Replies]

To: Lancey Howard; AFPhys

I need to learn how to take 10 years (heck, lets make it 30) years off my face/body digitally. Some people comment how good I look in pix. I mention that they are NOT close ups.. they are far way shots.

Then I need to apply it TO MY BODY w/o expense trauma of plastic surgery.

Now I am just positive the great FReeper Geeks here can help

lol!


24 posted on 06/05/2008 5:27:38 PM PDT by DollyCali (Don't tell GOD how big your storm is -- Tell the storm how B-I-G your God is!)
[ Post Reply | Private Reply | To 23 | View Replies]

To: yankeedame

thank you.. bookmarking!


25 posted on 06/05/2008 5:41:27 PM PDT by DollyCali (Don't tell GOD how big your storm is -- Tell the storm how B-I-G your God is!)
[ Post Reply | Private Reply | To 1 | View Replies]

To: jkphoto; Kitty Mittens; Mama_Bear

you all might enjoy this article/thread


26 posted on 06/05/2008 5:42:22 PM PDT by DollyCali (Don't tell GOD how big your storm is -- Tell the storm how B-I-G your God is!)
[ Post Reply | Private Reply | To 25 | View Replies]

To: DollyCali

Interesting. I am hard pressed to figure out the reasoning for people to doctor photos as they have. I’m especially naive or obtuse, I suppose.


27 posted on 06/06/2008 8:13:09 AM PDT by AFPhys ((.Praying for President Bush, our troops, their families, and all my American neighbors..))
[ Post Reply | Private Reply | To 24 | View Replies]

To: DollyCali
Thanks, Dolly. I also sent the article to JK (he rarely checks his pings). He said it would be interesting to discuss this with his photography class next fall.

This is a very amateurish example of faking a photo. This was found on a freeper's profile page several years ago. It's easy to spot the lighting problems and really bad cloning in this one.

There are many faked photos that are so well done that it is impossible to tell unless you know what you are looking for. It used to be, people would say "seeing is believing". Not anymore.

28 posted on 06/06/2008 8:36:47 AM PDT by Mama_Bear (My heroes wear camouflage!)
[ Post Reply | Private Reply | To 26 | View Replies]

To: JRios1968; DollyCali
You are correct, JRios1968. On the surface, at a casual glance, it looks okay, but a closer look shows an index finger that is obviously too long......

and, upon enlarging the pixels, many other poorly done areas show up....

Those are just two obvious areas. The lighting (particularly on his head and on his leg) is also a problem. Someone put a lot of work into this, but they didn't pay attention to some of the details.

I enjoy digital forensics - it's fun dissecting images. "Digital" has changed everything. A picture is still worth a thousand words, but the question now becomes, are they true. ;-)

29 posted on 06/06/2008 9:06:10 AM PDT by Mama_Bear (My heroes wear camouflage!)
[ Post Reply | Private Reply | To 9 | View Replies]

Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.

Free Republic
Browse · Search
General/Chat
Topics · Post Article

FreeRepublic, LLC, PO BOX 9771, FRESNO, CA 93794
FreeRepublic.com is powered by software copyright 2000-2008 John Robinson