Identifying Manipulated Images

Jamie found a cool story at MIT Tech Review. (As an aside, it sits behind an interstitial ad AND on 2 pages: normally I reject websites that do that, but it's a slow news day, so I'm letting it through.) Essentially, software is used to analyze light patterns in still photographs. Once you can figure out where the light sources are, it becomes a lot easier to determine if an image has been photoshopped.

Comments

Steganography

[unbug]

Does it also apply to steganography? Would sort of suck if it did.

Re:Steganography

[L4t3r4lu5]

*Somewhere in the middle of the NSA / MI6 buildings, a check mark is put next to an IP address.*

Re:Steganography

[JeanBaptiste]

Steganography is a good way to hide things from your Grandmother. I wouldn't trust it much past that.

Re:Steganography

[blueg3]

There's not much wrong with steganography of encrypted data, particularly if the data in the covert channel would have been statistically similar to random data anyway.

Most image steganography isn't that great, though, and steganography by a well-known means of cleartext data is fairly pointless.

Re:Steganography

[jd]

This would presumably work best with HDR images, such as JPEG2000, OpenEXR and the like, as very few image capturing devices could be reliable at 48bpp. Actually, the "ideal" might well be to calculate the degree of randomness in the low-order bits of the image and then pick an encryption algorithm whose typical degree of apparent randomness was closest to the degree of randomness in the image. There may be ways of using compression prior to encryption, or data recovery bits + deliberate corruption to further massage the encrypted data so that statistical treatment would find no abnormalities.

Re:Steganography

[Clete2]

Nikon has had a program out for a while now that can tell if a picture has been photoshopped. However, it only works on Nikon SLRs.

http://nikonusa.com/Find-Your-Nikon/Product/Imaging-Software/25738/Image%20Authentication%20Software.html

I wonder how the accuracy of the technique in the article is with a point-and-shoot?

Detector == Quality Control

[Anonymous Coward]

People who manipulate images will use these tools for quality control: When the fabrication passes all tests, it is ready to be released.

Re:Detector == Quality Control

[PrayerlessApostle]

Well very few will get released then. In that case the people who invented the detection technology have won then, haven't they? Fewer fakes released because the bar has been raised a lot now in terms of "quality control". Just because some photoshop nerd uses this for quality control doesn't mean it'll embue him with the power to photoshop pure gold fakes everytime he lays his hand on the mouse. If everyone implemented this as quality control before they released, there would be a lot less releases per week/month/year/whatever. Sure some people will still get through the net with fakes that fool this detection technology. But the point is very little will. So either a lot more fakes will get caught when this proliferates, or a lot more fakers will spend more time (a lot more) in photoshop trying to make an image that won't be found out to be a fake. A lot of them will probably never be able to release then, or will have to release with now-detectable faults in their image, because the bar is just too high now.

Re:Detector == Quality Control

[xquark]

If they can automate the detection then they can as easily automate the circumvention.

Re:Detector == Quality Control

[jd]

Detecting discrepencies between light sources is something that can be written as an algorithm and should increase in complexity linearly with the product of the number of light sources and number of pixels. Determining the correct value for pixels is something that must be written as a herustic and will increase in complexity exponentially with the number of objects (including light sources and non-visible objects) in the system. A herustic is not guaranteed to finish. Ever. Even if it always did, because the problem is exponential, far more images can be checked than can be fixed up.

Re:Detector == Quality Control

[ConceptJunkie]

You can automate the detection of fakes, but not the creation of fakes. That's the hard part. Besides, given the utter garbage quality of many faked pictures passed off as real, it's obvious a lot of people Just Don't Care to verify things.

Re:Detector == Quality Control

[Tjp($)pjT]

Once you can figure out where the light sources are, it becomes a lot easier to determine if an image has been photoshopped.

So use the new tool, determine the light sources and then add the correct light sources to you well rendered CGI model. Then the tool says it is "OK" and untampered with? Looks like a Photoshop plugin could be developed to check for this, and then maintain consistency. Just what we need. Better photo counterfeits!

Re:Detector == Quality Control

[chgros]

If they can automate the detection then they can as easily automate the circumvention.
It's easy to check a crypto signature but hard to generate it (if given only the public key).
I'm not saying there are any similarities between this and public key crypto, but just because you can check something easily, doesn't mean you can generate something that passes the check easily.

Re:Detector == Quality Control

[TempeTerra]

I don't follow your reasoning. Surely any heuristic for checking light values on a pixel will have to generate an approximation to check against. Fixing the image (relative to the checking heuristic at least) would just be a matter of setting the value in the image instead of checking it - so the complexity of checking and fixing images would always be equal if you know the checking algorithm.

Not sure what you mean about heuristics not being guaranteed to finish. A heuristic is just a quick-and-dirty approximation algorithm, many are written so that you can keep throwing processor cycles at them to get a progressively better answer, but if you expect to get the Right Answer it's not a heuristic any more.

Re:Detector == Quality Control

[jd]

An accurate method won't give you a value to test against, it'll give you a set of values for each pixel, where the value of that pixel contributes towards the values in the remaining sets for all the other pixels. You do not get a yes/no answer. The best you can do is a maybe/no. Simply put, let's just take the case of multi-generational diffuse reflections. Each object in the scene contributes something to the lighting of every other object, directly or indirectly. By adding or removing even a single object, the diffuse reflections for all pixels become wrong. You then have to calculate the inverse of a one-way function for each pixel, and then calculate the intersection of all the results of that, and then regenerate the scene, which gives you a really bad exponential problem. Whereas, if the photograph is complete, if you take enough pixels to develop an idea of the system, you can constrain the number of possible values of the remaining pixels sufficiently to be able to test if any given pixel is in or out of that set in very short order. It won't prove the image authentic, but it will be able to prove it inauthentic, if it is, on many occasions. How many depends on how constrained the system is. It can never be constrained to 1 possible solution, but you can certainly constrain it enough to have a high level of confidence in an image being genuine if no fakery is detected.

Re:Detector == Quality Control

[TempeTerra]

Ah yes, that makes sense. I was thinking of an over-simplified version of the problem. Thanks.

Re:Detector == Quality Control

[jd]

You've got me curious - what was the method you were thinking of? Particularly in noisy images, my method may produce an impractical set of sets of possible results, simply because the uncertainty is too high. In such a system, a simpler method that is less influenced by noise may be far more suitable.

Everything is photo-shopped!

[jimboindeutchland]

Duh!

Re:Everything is photo-shopped!

[ls -la]

Damn, you beat me to it.

Re:Everything is photo-shopped!

[mlush]

Damn, you beat me to it.

I'm sure there is an xkcd strip to cover this eventuality....

I'm sure its in here somewhere

Expert User Required

[Anonymous Coward]

TFA says an "expert user" is required. This expert user inputs coefficients that drive the equations that analyze the picture.

So basically, if you want an image to be doctored, you use one set of values. If you want an image to be genuine, you use another set of values. Maybe somebody else's requirements differ from mine, but this is not the kind of flexibility I want in a tool that is supposed to tell me if an image has been altered or not.

For an example of a better tool, see this article from Slashdot in August 2007.

Re:Expert User Required

[general_re]

TFA says an "expert user" is required. This expert user inputs coefficients that drive the equations that analyze the picture.

So basically, if you want an image to be doctored, you use one set of values. If you want an image to be genuine, you use another set of values. Maybe somebody else's requirements differ from mine, but this is not the kind of flexibility I want in a tool that is supposed to tell me if an image has been altered or not. Ummm, what? FTA:

Johnson's tool, which requires an expert user, works by modeling the lighting in the image based on clues garnered from various surfaces within the image. (It works best for images that contain surfaces of a fairly uniform color.) The user indicates the surface he wants to consider, and the program returns a set of coefficients to a complex equation that represents the surrounding lighting environment as a whole. That set of numbers can then be compared with results from other surfaces in the image. If the results fall outside a certain variance, the user can flag the image as possibly manipulated. I mean, that's not even close to what you posted - "running the same analysis on different parts of the image and then comparing the results" is not the same as "you pick the results".

Re:Expert User Required

[fredklein]

"running the same analysis on different parts of the image and then comparing the results" is not the same as "you pick the results".


It is if you get to pick the parts of the image.

I can tell...

[Anonymous Coward]

I can tell just by looking at the pixels and cause ive seen a lot of 'shops.

Re:I can tell...

[Anonymous Coward]

Damn. I wanted that one.

I for one welcome our new manipulated image sensing overlords.

No ads, all on one page

[wile_e_wonka]

The printer-friendly version:

http://www.technologyreview.com/printer_friendly_article.aspx?id=20423

Re:No ads, all on one page

[IBBoard]

Clever (yet annoying). I found that link as well, but it doesn't work.

First time I hit the article (from the link in the summary) it loaded fine. Clicked your link and I got an ad with the "skip this ad" link (although the ad wasn't there because of AdBlock) then it took me to the main article across two pages.

Looks like those damned evil news people don't want us to avoid their adverts.

Re:No ads, all on one page

[wile_e_wonka]

Sorry people--it doesn't work.

Darn advertisers!

Re:No ads, all on one page

[trolltalk.com]

The link works fine if, instead of clicking on it, you cut-n-paste it into a new browser tab. Here's what you get, if you can't be arsed to go to the trouble:

Monday, March 17, 2008
Identifying Manipulated Images
New tools that analyze the lighting in images help spot tampering.
By Erica Naone

Photo-editing software gets more sophisticated all the time, allowing users to alter pictures in ways both fun and fraudulent. Last month, for example, a photo of Tibetan antelope roaming alongside a high-speed train was revealed to be a fake, according to the Wall Street Journal, after having been published by China's state-run news agency. Researchers are working on a variety of digital forensics tools, including those that analyze the lighting in an image, in hopes of making it easier to catch such manipulations.

Tools that analyze lighting are particularly useful because "lighting is hard to fake" without leaving a trace, says Micah Kimo Johnson, a researcher in the brain- and cognitive-sciences department at MIT, whose work includes designing tools for digital forensics. As a result, even frauds that look good to the naked eye are likely to contain inconsistencies that can be picked up by software.

Many fraudulent images are created by combining parts of two or more photographs into a single image. When the parts are combined, the combination can sometimes be spotted by variations in the lighting conditions within the image. An observant person might notice such variations, Johnson says; however, "people are pretty insensitive to lighting." Software tools are useful, he says, because they can help quantify lighting irregularities--they can give solid information during evaluations of images submitted as evidence in court, for example--and because they can analyze more complicated lighting conditions than the human eye can. Johnson notes that in many indoor environments, there are dozens of light sources, including lightbulbs and windows. Each light source contributes to the complexity of the overall lighting in the image.

Johnson's tool, which requires an expert user, works by modeling the lighting in the image based on clues garnered from various surfaces within the image. (It works best for images that contain surfaces of a fairly uniform color.) The user indicates the surface he wants to consider, and the program returns a set of coefficients to a complex equation that represents the surrounding lighting environment as a whole. That set of numbers can then be compared with results from other surfaces in the image. If the results fall outside a certain variance, the user can flag the image as possibly manipulated.

Hany Farid, a professor of computer science at Dartmouth College, who collaborated with Johnson in designing the tool and is a leader in the field of digital forensics, says that "for tampering, there's no silver button." Different manipulations will be spotted by different tools, he points out. As a result, Farid says, there's a need for a variety of tools that can help experts detect manipulated images and can give a solid rationale for why those images have been flagged.

Neal Krawetz, who owns a computer consulting firm called Hacker Factor, presented his own image-analysis tools last month at the Black Hat 2008 conference in Washington, DC. Among his tools was one that looks for the light direction in an image. The tool focuses on an individual pixel and finds the lightest of the surrounding pixels. It assumes that light is coming from that direction, and it processes the image according to that assumption, color-coding it based on light sources. While the results are noisy, Krawetz says, they can be used to spot disparities in lighting. He says that his tool, which has not been peer-reviewed, is meant as an aid for average people who want to consider whether an image has been manipulated--for example, people curious about content that they find online.

Cynthia Baron, associate director of digital media programs at N

Re:No ads, all on one page

[Anonymous Coward]

Johnson's tool, which requires an expert user Yeah, my tool requires an expert user as well.

That should help

[WormholeFiend]

with all those UFO hoax photos then... /I want to disbelieve

Re:That should help

[zotz]

Funny thing, you don't always have to shop things to get odd results:

My vids on youtube:

http://www.youtube.com/user/zotzbro

If you check the comments on the "UFO vs Paper plane test" you will see people talking of a real one.

Perhaps on some of the paper plane instruction vids too. If you watch those, as the camera pans in one of them, after the construction and before the flight test, you can see what the "UFO" really is.

all the best,

drew
http://zotzbro.blogspot.com/

Re:That should help

[sm62704]

It's not needed and won't help. Most of the UFO photos are pre-Photoshop and were done with different methods:

• Have a small model of the UFO and fling it into the air high enough that there's no context. Although those CAN be detected, they can't by this software.
  
• The objects are secret military aircraft, not alien craft. The hoax of alien craft is started by the government (pick one) to mask the true meaning of the object photoed. This software won't help with that, either
  
• It's something else flying around up there. Is it a bird? Is it a plane? Is it a weather balloon? Is it ball lightning? Who knows? If it's a flying thing and you don't know what it is, then it's an Unidentified Flying Object. This tool won't help here, either.

This tool can't do anything someone trained in art can't do. The first thing you learn in art school is how to see. You can't draw if you can't see, and that's usually the biggest reason most people can't draw.

As one of my instructors used to say, "I don't know what I like but I know what art is."

-mcgrew

Re:That should help

[Paradise Pete]

My vids on youtube:

So why is your watch on your right wrist, and yet you throw right handed? I'm thinking the real secret is that you must be a photoshop creation.

Re:That should help

[Anonymous Coward]

It's something else flying around up there. Is it a bird? Is it a plane? Is it a weather balloon? Is it ball lightning?

Why it's superman of course! He's gonna descent upon you and bitch-slap your unbelieving rear end!

Re:That should help

[zotz]

"So why is your watch on your right wrist, and yet you throw right handed? I'm thinking the real secret is that you must be a photoshop creation."

Not sure, it feels better there. My dad was the same way iirc.

Supposedly, I was ambi as a kid until pre-school made me go righty.

I water ski like a lefty. I wind a top like a lefty but spin it from my right hand. I have a stronger kick from my left foot but am more accurate with my right... And a number of other oddities...

Unless I am a photoshop creation... wwnnsnmsnm...

all the best,

drew
http://zotzbro.blogspot.com/

Re:That should help

[Paradise Pete]

You're the first righty I know of to do that. I wear it on my right wrist, but that's because I'm left-handed. Even so some people still think it's weird. I'm ~100% lefty. The only thing I do right-handed is use scissors.

Colombo did this on his 1970's TV show

[jimwelch]

Someone wore a photo mask and tripped a speed camera to give their partner proof that they were across town (LA) at the time of the murder. He noticed the shadow under the nose was wrong by comparing previous and following pictures from the same camera.
I am not sure which episode it was. Peter Falk as Det. Lt. Colombo

Re:Colombo did this on his 1970's TV show

[R2.0]

Speed cameras in the 70's? Sure about that?

Re:Colombo did this on his 1970's TV show

[jimwelch]

The show ran from 1971 to 2003, like I said, I don't remember which episode, but I seen it several times.

Uh Oh

[Missing_dc]

This bodes ill for all those geeks out there with "out-of-state" girlfriends!!

Goes both ways

[Nerdposeur]

...and then the photoshoppers will write evolutionary algorithms to modify their photographs until they pass evaluation by this tool.

Re:Goes both ways

[Linker3000]

Yeah, but do they blend?

weak

[gnudutch]

this method is way better

Forensic Analysis Reveals Al-Qaeda's Image Doctoring

Re:weak

[SCHecklerX]

If you read the linked article, you'll notice that Neal has some responses to questions posted. He used both methods in his Blackhat presentation.

Limited utillity

[johnjaydk]

In a studio or other arranged settings it's pretty standard to use multiple lighting sources. So this tool will mainly be usefull for outdoor settings. If it's up-close and personal then it's also very common to use lights or other tools outside. Sooo this tool should be used with moderation.

Re:Limited utillity

[aadvancedGIR]

Moreover, it is well known that photoshop is a standard and commonly used tool for professional studio photography anyway. I think the tool purpose is limited to check that a "genuine" photography used to prove a crime or the existence of UFO/Bigfoot is not a blatant fake.

Adds a step for the photoshoppers

[crowemojo]

One would think that it would be simple enough, after finishing whatever touch-ups that you want to perform, that you use this technique to calculate where the light sources should be, and then correct the minute details that would give it away as an altered image. Sounds like the kind of thing that would be a simple photoshop plugin actually, once you are all done you just run the "make undetectable from light source detection analysis" tool and call it a day.

Re:Adds a step for the photoshoppers

[CambodiaSam]

Light sources have always been a pain for me when photoshopping. I'm not surprised that it's the key to this software, as it tends to be the most difficult aspect of manipulating an image (I'm a slightly more than casual user, but not a graphic designer). Light completely changes the color structure and I end up spending an inordinate amount of time trying to redo hues and fix shadows that don't line up.

If there's a plugin for helping me with that part of the struggle, I hereby scream to my fellow slashdotters to please fill me in!

Re:Adds a step for the photoshoppers

[Anonymous Coward]

This algorithm at least requires human intervention, to find contours (RTFA). Additionally, it's easier to find inconsistencies than to automatically resolve them in a way which doesn't cause potentially something else to look weird.

It's just another tool which requires (some) human ingenuity to use effectively and thus requires (just a little more) human ingenuity to use offensively. After all, you only need to miss one detail to be found out... and hiring a photodoctoring expert is something you don't want to do if you (e.g. news media) are trying to sneak this past the public. I'd say that it's not as symmetric as you suggest.

You can come up with a lot of confounding examples for this method anyway - imagine if one of the subjects in your photo has a flashlight, or is illuminated by a spotlight. "Make undetectable from light source analysis" (or "Detect Forgery by Light Source Analysis" for that matter) is something I'll expect to see in CSI, not reality.

Finding Photoshopped Pics for Fun

[Dan East]

Does anyone else have a habit at looking at pictures and trying to see how they've been manipulated? These types of pictures are rampant in advertising. Pick up any magazine and start looking, and the poorly edited pictures will jump out quickly. The more professionally edited pictures have much more subtle problems, and can take a bit of poring over to find. Many product images (on packaging and in catalogs) are the same way, and are usually the worst edited of the bunch. Some things I look for:

• An object rubber-stamped in multiple places. Each copy is identical, which gives it away. They are often scaled, rotated or mirrored to make them look more unique.
  
• Lighting and shadows, which is what the algorithm in this story deals with specifically.
  
• Focus. Often multiple objects will be in focus at varying distances impossible with a single shot.
  
• The same image of a person is used in multiple shots. This is most prevalent in product images in catalogs.
  
• Poor masking, where edges of objects are over or under processed, either clipping part of the object (hair can be particularly tough to do), or showing some color edges from the original background.
  

Anyway, that's just the geek in me I guess, because I really do enjoy finding flaws in images. What I hate is an image that has a sort of surreal perfection to it that I know must be composited, but I can't find any smoking gun.

Re:Finding Photoshopped Pics for Fun

[jo42]

"If it looks too good to be true, it is."

Re:Finding Photoshopped Pics for Fun

[Dmala]

Playboy is probably the worst offender at this. Most of the women these days are so heavily airbrushed/Photoshopped that they look more like paintings or cartoons than actual photos.

Er... I mean... I just read the articles, but that's what a friend told me about the pictures.

Re:Finding Photoshopped Pics for Fun

[hairykrishna]

Essentially all catalog images are manipulated. Some are even computer renders rather than real photos. My housemate used to work for a company that imported a lot of the stuff sold in Argos (UK crappy catalog chain). They used to sometimes submit the product photos before the actual items existed, while they were still being manufactured in China.

Re:Finding Photoshopped Pics for Fun

[Atario]

Focus. Often multiple objects will be in focus at varying distances impossible with a single shot.

This is perfectly possible, given a small enough aperture setting in your camera and correspondingly high light level in your studio.

Re:Finding Photoshopped Pics for Fun

[Dogtanian]

Some are even computer renders rather than real photos. Yeah, but doesn't the "cheap plasticky" appearance of the rendered product shots give it away? Oh, hang on....

a lot of the stuff sold in Argos ....scrap that question. In Argos' case, I doubt anyone would be able to tell the difference! ;-)

Re:Finding Photoshopped Pics for Fun

[Blue Stone]

>"Focus. Often multiple objects will be in focus at varying distances impossible with a single shot."

Not at small apertures (high F-stops).

Re:Finding Photoshopped Pics for Fun

[DNS-and-BIND]

Ah, yeah...no kidding. I work in China and view a lot of factory catalogs, they have lots of cheap manipulations like that. The creepiest one was a catalog for a garment manufacturer...on every page, on every outfit, was the same grinning blonde guy. After a few pages, it started to get funny. After a few more pages, it got geniunely eerie. Every page, the same rictus grin, flipped left and right. I had to put the catalog down after a while, it was getting to me. Sort of like Aphex Twin's "Come to Daddy" video.

Good

[Fri13]

Now when we have a tool what shows when image is manipulated by using Photoshop, we can start using GIMP or any other _image manipulation_ software because those tools cannot trace them because they dont "photoshop" images, they manipulate them.

Yah, bad sarcasm, im just tired that "photoshop this" "photoshop that" like there would not be any other image manipulation software. I bet that over 50% Photoshop owners just has a warez version of it and 80% of photoshop users could do their things with any other software.

Based on the same codebase, in fact

[phorm]

You could even use the same codebase. If a particular method is using light-sources/shadow/etc to determine the authenticity of an image, then you could plot those same light-sources and have a plugin modify the image to be what the authentication plugin would expect.

Re:Based on the same codebase, in fact

[Neotrantor]

horse shit

changing the light source in a picture requires that you might have to desaturated some pixels and guess what their level/color might be, which is information you probably don't have available.

Content specific test

[grikdog]

Does the "image" contain Lindsey Lohan, Paris Hilton, Hillary Clinton or Sarah Michelle Gellar? It's photochopped.

Re:Content specific test

[grikdog]

Actually, I used to think the puffs of "white smoke" and "black smoke" in the Challenger disaster films were early photoshops. A puff of zeroes. A puff of ones. Recently, though, NASA seems to have released extraordinarily detailed movies (from the classified cameras and undocumented camera angles, evidently) that do show beautifully complex puffs of thick black roil. Maybe they can do that with fractals these days, but if so, is it still Photoshop? GIMP can't do that (yet).

Apollo

[sir_eccles]

I love how the first comment is asking if the apollo landing photos were photoshopped.

Re:Apollo

[lena_10326]

I love how the first comment is asking if the apollo landing photos were photoshopped.

Photoshop was HUGE in 69. Huge.

That must be why

[sleeponthemic]

Elvis is always spotted in 7-11.

Easily masked light sources (hint: they're everywhere!).

Ad? What ad?

[brunes69]

Adblock - everyone should have this installed.

Re:Ad? What ad?

[francisstp]

Or use Opera which blocks these by itself...

Re:Good

[Marvin01]

It just sounds wrong to say that an image has been "GIMPed".

Actually, now that I think about it, I kinda like it...

Found the episode

[jimwelch]

Columbo and the Murder of a Rock Star, 1991 and yes Dabney Coleman was the bad guy.

Re:Found the episode

[boredsenseless]

Wasn't Dabney Coleman always the bad guy?

Re:Good

[qoncept]

they dont "photoshop" images, they manipulate them.

I hope you don't have a runny nose or a paper cut, because you'd probably get upset if someone offered you a Kleenex or Bandaid. Unless those don't fit your agenda. All the pros are using MS Paint anyway.

Re:Good

[glwtta]

im just tired that "photoshop this" "photoshop that" like there would not be any other image manipulation software

Do you also hang around the Epson at work explaining to people how they aren't really "xeroxing" anything?

TFA with no ads and on one page

[ErichTheWebGuy]

http://www.technologyreview.com/printer_friendly_article.aspx?id=20423 Enjoy :)

Well, there goes the pr0n industry...

[amplt1337]

nt

The opposite

[kcdoodle]

I have written code that does the opposite.

I have shape generating program (wire frame sphere distortions) I developed. I wanted to add shading to the shapes. So I pick a point for the light source at random and color the pixels closer to the light a little brighter than those pixels farther away from the light. Really works well.

Once I got that working, I wanted to add self-cast shadows. You know, when the generated shape twists back on itself and should cast a shadow onto itself. Never got there, still sits in my code archives as just a "cool dealie".

Shouldn't be too hard to work my code backward and calculate (or guess) where lighting is coming from. I mean heck, I am one guy and this isn't so hard.

Horse shit ^ horse shit

[hummassa]

Sorry for the subject line, but...
if you have enough information to test it, then you have enough information to fake it (at least well enough to pass the test).

I don't see how this is useful.

[Tokimasa]

If you analyzed the noise pattern of the image, any edits would be obvious - the noise pattern wouldn't exist over edited components.

Re:I don't see how this is useful.

[argent]

That's another approach, how well does it do when the manipulator layered noise over the image.

Re:I don't see how this is useful.

[icegreentea]

Would printing and then scanning 'normalize' the noise? I'm curious.

Re:I don't see how this is useful.

[plover]

That depends greatly on the quality of the printer, the quality of the scanner, and the quality of the scanning job. Higher resolution printers and higher resolution scanners may actually reproduce the noise accurately enough to be picked up. A Xerox "watermarking" method proposed in the 1990s consisted of color regions being rendered by tiny diagonally oriented color bars. These were designed specifically to survive photocopying and scanning.

I'd guess that modern printers and scanners would probably not 'normalize' the noise at their default settings. You'd probably be more successful if you tried to digitally correct the noise. Resolution changing, format changing (GIF to JPG), compression (higher JPG) all will somewhat smooth out some of the artifacts, but they are certainly not guaranteed protection against forensic analysis.

Re:I don't see how this is useful.

[argent]

I was thinking more along the lines of rendering white or pink noise directly onto the image, either directly or by applying a randomized convolution to the whole thing.

Re:Good

[dotancohen]

That's why I say "to gimp a photo" rather than say "to photoshop a photo". It spreads awareness, breaks the Adobe monopoly, and sounds more natural. Even Adobe discourages the use of "photoshop" as a verb.

Faked non-uniformly lit

[abqaussie]

The tool doesn't tell you if a photo is faked, it just analyzes whether there are light sources in the image that are not affecting different objects in the image the same way. From what I can tell it tries to tell if the way the light hits different objects in the picture "agree" with one another based on the position of the object, color, and probably other attributes not detailed in the article. If the photographer is controlling the light at all, using off-camera flash, focusing their light on some parts and blocking it from others, etc, then there would be components of the image that deliberately don't match when it comes to the lighting. People do that all the time, both deliberately and accidentally, when lighting a photo. Because the photographer has deliberately put a light on the subject that isn't hitting other elements, background, objects, the same way as it's hitting the subject. So it seems like the analysis would work great for cases where the light is ambient, and should affect all objects in the frame relatively the same. Otherwise it'd have a bad day.

Re:Good

[Mister Whirly]

"im just tired that "photoshop this" "photoshop that""

Just "Google" it. It is a legitimate verb now.

"I bet that over 50% Photoshop owners just has a warez version of it and 80% of photoshop users could do their things with any other software"

And I bet there is a 100% chance you pulled those numbers from right out of yer ass. Thinking something doesn't automagically make it a fact. And haven't you ever seen the GIMP vs. Photoshop discussions like a million times on Slashdot before? If not let me condense it for you - GIMP works great as a free software tool, but most professionals wouldn't use it to do their paid work, as it lacks some of the higher end features of Photoshop.

Re:Faked vs. non-uniformly lit

[abqaussie]

Title on previous post should have read "Faked vs. non-uniformly lit"

Re:Good

[m.ducharme]

Even Adobe discourages the use of "photoshop" as a verb. You know why, right? If they let "photoshop" be corrupted in the language as a verb, they would eventually lose the trademark rights to the name. Eventually, companies would be able to get away with naming their software "MS Photoshopping Program" or "Gimp Photoshop Utility" or whatever, and Adobe wouldn't be able to do anything about it.

Are you sure now, that you want to discourage people from using "photoshop" as a verb?

Donnie Hoyle is right...

[actionbastard]

You all suck at Photoshop.

No, really, you do.

NASA was here

[brainwash]

Time to run the test on the Apollo shots, for all the tinfoil hats.

Re:NASA was here

[Single GNU Theory]

No amount of proof is going to convince the true nutters, tinfoil hats, and conspiracy theorists. They're very focused, and the ones I saw on a recent documentary about the ones who claim the moon landings were faked (on either The Discovery Channel or the National Geographic Channel, I can't remember which) seemed to be the kinds of guys who have a lot of time on their hands. No mere truth would be enough to persuade them to believe differently.

strings

[TheDarkener]

Or, with most dumb$*@&s who like to play off photoshopped images as real,

$ strings pic.jpg | grep -i photoshop

Re:strings

[blueg3]

It's a lot easier to edit those fields than to properly redo all the lighting so that a lighting-analysis program can't detect it.

anyione

[SnarfQuest]

Anyione else notice the tyipo?

speaking of which

[circletimessquare]

did anyone else catch the blog in the new york times about the fenton photographs?

apparently this guy took some photos of some cannonballs in the crimean war that became famous as a poetic commentary on war. this documentary filmmaker, errol morris, has gone completely unhinged obsessive compulsive over whether or not the photos are fake and/ or manipulated. it's utterly fascinating, and a little weird, to see so much time and effort devoted to these photos. specifically, cannons and shadows. utterly esoteric and thorough. he also expands into the larger topic of the history of manipulated politically sensitive photos. makes for a good read, especially if you are interested in pre-photoshop image manipulation

check it out, talk about thorough

Re:speaking of which

[argent]

There appear to be a number of balls in the "OFF" picture that do not appear in the "ON" picture, further up the hill than the ditch. To make it easier to find the differences I aligned the images and applied a false color mask using red and green overlays for the two images:

http://scarydevil.com/~peter/images/Extra-balls-in-OFF.png
http://scarydevil.com/~peter/images/OFF-plus-false-color.png
http://scarydevil.com/~peter/images/ON-plus-false-color.png

Silver Button?

[Jaqenn]

Hany Farid, a professor of computer science at Dartmouth College...says that "for tampering, there's no silver button." Silver button? What the heck is a silver button?

I'm assuming he was trying to say silver bullet. Do we blame the professor, or the journalist?

Re:Silver Button?

[Anonymous Coward]

No, a silver button is a well known thing...made with Photoshop and frequently used on websites as a thing to click on.

What is "manipulated?"

[toddhisattva]

How are we to define "manipulation?"

Gamma adjustment and color correction. Dodging and burning. Red-eye and dust removal. Cropping. JPEG or other data compression. Dynamic range compression.

The only non-manipulated image is the raw data from the sensors. This is actually an improvement over film, where developing techniques can have quite an impact on the negatives - in film photography, there are no un-manipulated images.

Any serious news bureau should provide their raw images to interested parties. Only this way can their customers (us!) tell how much manipulation an image has undergone.

Analyze, not "analyize"

[1u3hr]

Taco, I will try not to make the obvious "anal" jokes, but FFS, is clicking spellcheck really so hard?

Re:Analyze, not "analyize"

[Hatta]

Thanks for rectifying that error.

Re:Good

[Anonymous Coward]

and 80% of photoshop users could do their things with any other software.

Assuming this number even vaguely reflects reality, you're right; they could. But they don't.

So what you're telling me

[Provocateur]

...is that Keith Richards actually looks like that? Without any enhancements at all?

Some other stuff

[SCHecklerX]

Light source analysis was one of several methods used at a talk at Blackhat DC this year. The much more visually impressive tool, for me, was the ability to show quite explicitly what has been modified in a lossy-compressed (like jpeg) image:

http://www.blackhat.com/presentations/bh-dc-08/Krawetz/Presentation/bh-dc-08-krawetz.pdf

Compresion analysis tool:
http://www.blackhat.com/presentations/bh-dc-08/Krawetz/Extra/jpegquality.c

Re:Good

[Anonymous Coward]

no they wouldn't, everyone calls vacuum cleaners hoovers, doesn't mean that Electrolux can sell vacuum cleaners with hoover written on them.

"photoshopped"?

[Lxy]

When did "photoshop" become a verb?

This post has been gimped by the gimper

Re:"photoshopped"?

[Opie812]

In 1990. Where have you been?

Re:"photoshopped"?

[Anonymous Coward]

probably out nit picking some other random nonsense. he must get a real rise out of it.

Re:"photoshopped"?

[HeroreV]

You could probably find out if you googled it.

Re:"photoshopped"?

[Lxy]

no time for that, I've tasked one of my coworkers to check into it.

Another good tool for detecting photoshopping

[Reziac]

...and for diagnosing damaged JPGs (I used it extensively when reconstructing mangled JPGs from someone's disk crash):

JPEGsnoop, by Calvin Hass
In very active development; suggestions and bug reports welcome. Free download from http://www.impulseadventure.com/photo/jpeg-snoop.html

way better?

[CrazeeCracker]

The interesting thing about that is... The code used in the article the OP linked to features the following lines:

Revision history: This code has been stripped out of imgana by Hacker Factor Solutions. (Imgana does much more than quality analysis, but that's all that is being released right now.) Said program by Hacker Factor is also mentioned in TFA as a more basic approach to checking whether or not an image has been manipulated. I'll leave you to judge what this means.

As an unrelated sidenote, Hacker Factor features a very interesting javascript that guesses the gender of the author of a block of text (>300 words). Thus far, I've found it to be eerily accurate.

Re:way better?

[cleatsupkeep]

What do you think it would say about this? (bash.org)

(JHawk111420) Hey whats up, a/s/l?
(Lady Renegade) more than you want, I'm sure :)
(JHawk111420) ill take that as a challenge ;-)
(Lady Renegade) take it any way you want sweetie
(JHawk111420) k, how old are ya?
(Lady Renegade) probably too old for you, but let's pretend I'm 20 ;)
(JHawk111420) k, what do ya look like?
(Lady Renegade) before or after I'm dressed up?
(JHawk111420) both :-D
(Lady Renegade) well......after I'm dressed up, I have long sexy red hair, nails painted red to match the slinky dress I have on, stiletto heels, pouty lips, green eyes, boobs out to here, and a smile that stops
traffic
(JHawk111420) and before your dressed up?
(Lady Renegade) before I'm dressed up, I'm bald and wearing boxers...sometimes my weenie is peeking out
(Lady Renegade) hello?
(Lady Renegade) hello?
(Lady Renegade) hello ....

Re:way better?

[Peter Mork]

I haven't tested it extensively, but every chunk of text I threw at the program, it asserted was authored by a male. And yet, more than half the samples were from papers/blogs/fiction authored by female friends. I think that the program assumes that if you are even vaguely literate, you must be male.

dude

[circletimessquare]

send your work to errol morris. he is sure to use the data and give you a shout out

Re:dude

[Reziac]

I didn't go that far. I looked at the length of the road visible beyond the foreground area, concluded that it had already been cleared of loose cannonballs (since there are none on the rest of its visible length), and that therefore the cannonballs lying on the road had to have been placed there FOR the 2nd photo.

Also, the area appears to have been slightly muddy, and the cannonballs on the road show no sinkage, while those elsewhere do. Ergo, those on the road have been there less time than those in the ditch.

Re:Good

[Herkum01]

You photoshopped my comment!

Re:Good

[Reziac]

Explains some of the lame images we see on the net ... ;)

Re:Good

[Reziac]

Actually, they have to be SEEN to discourage its use, lest they lose the trademark.

But you can bet Adobe is secretly pleased that almost every time someone mentions a manipulated image, their brand-awareness gets a small but definite boost. I know people who think that ONLY Photoshop can be used for such manipulation, in part thanks to this verbification.

Re:Good

[zippthorne]

Indeed. The number is probably closer to 90% of photoshop "owners" have warez copies. At least based on my own, rather poor sampling: I have met two, maybe three people who have or had at one point warez copies of photoshop, and only zero people who are actually professional graphic artists. Therefore illicit casual users vastly outnumber professionals.

Re:Good

[blincoln]

I'm sure it depends on what community your data is sampled from.

In the business world, most copies are probably legitimate. Hobbyists and amateurs probably have bootleg copies (aside from the serious hobbyists who feel compelled to buy one), because Adobe prices it as a business tool. Plain old Photoshop CS3 is US$650 for the full version. "Extended" is US$999. If you want the whole Creative Suite, it ranges from US$1199 to US$2499. I do photography as a hobby, and my camera (a secondhand D70) cost less than $650.

Re:Good

[m.ducharme]

True. Of course, it's a somewhat dangerous game to play, actually losing the trademark would be a big fail for Adobe. The resulting flood of ersatz photoshop products would wash the marginal gains in brand awareness away.

Also, Adobe's brand would be eroded, not marginally boosted, if there were actually a feature-for-feature competitor for it's product. Note how kleenex is such a generic term, because there is no real difference between a Kleenex, and any other brand of paper snot-wipes.

Re:Good

[The Queen]

So is Google discouraging this practice, as well? Seems a bit late.

You read the links?

[adavidw]

... normally I reject websites that do that...

Sorry, Taco, I'm not buying it. This statement implies that Slashdot "editors" actually read the links in the submitted articles, or at least click on them. We all know from past experience that that's just not true.

Duh

[certsoft]

That's how Lieutenant Gaeta cleared Dr. Baltar over the faked photograph.

Re:Good

[m.ducharme]

Yes, they are (or rather, they have). Google sent out a round of letters in 2003 and 2006 to various websites where "google" was being used as a verb. Google "google trademark protection" for examples.

Trust in your source is more important

[photomonkey]

Of course, we've heard stories about staffers at papers and magazines faking or 'enhancing' photos. If it happens with people ON STAFF, then why do news outlets take hand-out art from companies, foreign governments and other non-trusted sources?

For years, I have made my living as a freelance news photographer, and am a member in good standing with several professional organizations. Sure, I could still lie, doctor photos or submit work that isn't my own; but I have added incentive not to: I depend on being trusted to sell my work. If I can't be trusted, I can't get jobs. So beyond my personal ethics paying respect for professional ethics, I have financial incentive not to make shit up.

But alas, as newsrooms shrink their staffs and call out for less freelance work, my as-true-as-I-can-humanly-make-them photographs will be left to go by the wayside in favor of hand-outs from people and organizations who have no incentive to follow the ethical rules AND an interest in making themselves look good.

I guess that's why I find myself photographing more weddings and corporate/advertising pieces. Oddly enough, they are clients who EXPECT to be lied to with heavily retouched photos.

And for the record, it is my professional opinion that this software won't make a bit of difference.

Any of you guys work for international media companies that are looking for America-based freelance contributors? I know a lot of the overseas media still looks for truth in reporting.

I found it eerily inaccurate

[tlambert]

I found it eerily inaccurate.

I threw a bunch of random samplings of English text from Project Gutenberg at it. It claims almost everyone is male.

-- Terry

...I can tell by some of the pixels

[toby]

And having seen quite a few 'shops in my time...

Re:Good

[tepples]

That's why I say "to gimp a photo" rather than say "to photoshop a photo". It spreads awareness At least one YTMND user agrees: "GIMP up an image, make it look bad".

Re:Good

[m.ducharme]

I don't know anyone who calls their vacuum a hoover. I call mine a roomba. Anyway, it may be that Hoover's brand isn't dilute enough, or it just may be that no company has ever tried it, but it does happen that companies do lose some or all of their trademark rights if their brand names become common coin.

Re:Good

[plover]

Not that all vacuum cleaners were called "hoovers", but your parents or grandparents may have hoovered the rugs more than once in their lifetimes. It's also possible that "to hoover" was regional slang, perhaps more so here in the Midwest. Either way, it's definitely fallen into disuse over time, and never stuck like kleenex or xerox.

Voodoo software

[Anonymous Coward]

The "tool" described in the article is a joke and reminds me of one (that also got posted on Slashdot, some time ago) whose author claimed to be able to tell if an image was a composite of several sources by "analyzing the compression patterns". Turned out all his tool did was check the quantization level of JPEG macroblocks... which is always defined by the last encoding. In other words, all it did was detect which parts the (last) encoder has decided to compress more, and which parts it had decided to compress less. Basically it worked like a (really bad) edge detection filter.

The description of the "tool" in this article is just as nonsensical. If a pixel is brighter than its neighbor, that tells you nothing about the position of the light source. If you want to determine that, you need to do some very complex spatial analysis, like the stuff done by Paul Debevec (and it's still only an approximation).

Human eyes and brains are much better at spotting lighting inconsistencies than any algorithm, because they can fill in 3D information from 2D images (which you simply cannot do without a powerful interpretative visual system). This tool (which, as the article states "has not been peer-reviewed" - what a surprise!) is just a hack put together to impress some wannabe "security consultants" and get a fat check for a quick patent buy-out.

Re:Good

[SteveDob]

OK, I'm a UK guy in his mid 40s, so the subject doesn't come up all that often. This is obviously anecdotal, but I don't know of anyone who *doesn't* call it a hoover. Possibly this is a regional/generational thing. Mine isn't even a Hoover, it's a Dyson, but there is no way I would ever consider saying that I'm dysoning the carpet.

Re:Good

[dotancohen]

Nice, thanks, too bad I don't have modpoints. I get the feeling that the page was written to catch certain keywords, for those folks googling around.

One good thing out of this

[Xen2010]

At least it will hopefully put a stop to the horrendous movie posters I'm seeing like "The Accidental Husband" and "Over Her Dead Body". The photoshop job in these are SO BAD it makes me physically ill to look at them.

Photoshop & Verb

[Acer500]

This already happened in my country (Uruguay) for other stuff: we call chewing gum "chicles" after the Addam's Chiclets brand , and running/tennis shoes "championes" after the Champion brand (that term's unique to Uruguay I think, in Argentina they use the term "zapatillas").

Re:Good

[Reziac]

Yeah, it's a risky game all right. But Hormel seems to do okay with SPAM vs spam :)

Unfortunately trying to maintain a lost brand trademark by being the best in the field doesn't work, since most people buy cheapest regardless of quality. :(