Identifying Manipulated Images

Identifying Manipulated Images

www.technologyreview.com

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.

(visit the link for the full news article)

I found this in one of the comments posted on the site.

"I would be nice to use this tool to determine if any of the Apollo moon landing photos were faked. Theoretically there should only be a single light source, shouldn't there?"

I thought that was brilliant, and a excellent way to test this software.

They never did mention where to get the software only that you had to be an expert to use it.

I can't wait for a photoshop plug-in!

But seriously wouldn't this put the moon conspiracy to rest?



www.technologyreview.com
(visit the link for the full news article)

Thanks for the post, IMAdamn!

The more sophisticated software becomes, enabling the stealthy alteration of images, the better the tools evolve to detect the attempts.

But they still can't undo the alterations.

In other words, they can't remove the antelope and restore the image to it's pre-altered state.

All they can do is determine the image has been compromised.


Edit to fix teh grammar.

[edit on 17-3-2008 by goosdawg]

reply to post by goosdawg


Your right!

I would love to see this tool analyze the Apollo photographs.

Who cares if you can restore the original image in this case, we just want to prove a single light source, or evidence of a staged scene.

Thanks for the reply!

Sounds pretty cool. I know an expert can still spot a fake through examination alone... but who has the time to sit and analyze photos for 10 minutes at a time. Software to spot fakes is definitely a good idea.

I've seen alot of awesome pictures that I found were fakes... just normal seeming photography of people or animals doing something cool. What's funny is that while a picture of an Alien will cause everyone to scream fake, a faked picture of something normal like a person climbing a building is never going to receive such scrutiny. It's easy to fake images if the subject is already familiar to the viewer.

On a conspiratorial note, I wonder how many images are faked for political or judicial purposes?

If you're the cops, and you're sure that Joe Shmoe is the guy holding up liquor stores around town, can you fake a store video to use as evidence? You're the CIA and you want to make sure Atta is IDed as one of the hijackers, do you fake the picture of him going through the turnstile? How many political careers are effected, even of people in other countries, accepting a bribe or some other sting?

And do people stop and look at pictures on MSM and wonder how many are faked?

Food for thought.

If someone u2u's me the name of the proggie . I can probably get it . Just need whatever info can be had on it .

Originally posted by NGC2736
On a conspiratorial note, I wonder how many images are faked for political or judicial purposes?

If you're the cops, and you're sure that Joe Shmoe is the guy holding up liquor stores around town, can you fake a store video to use as evidence?

Great point!

If you think about it, surveillance cameras throughout the world are far from a fluid video. When was the last time you have seen a fluid motion security cam? My guess is never. They compress the image and drop frames to save room on the hard drive or video media.

All you would have to do is splice a couple of frames(pictures) within the video and come up with any evidence you wish.

Scary to think about, but, did we all not already know this?

I want this tool!


When does conspiracy become "fact"?

This is a fun little test you can take: Fake or Photo?

I got every single one right, excluding one image in the bonus round. But I've been using photoshop for years, and can spot a fake pretty easily. Yes, the main thing is lighting. It is very very hard to replicate, I would say almost impossible without proper 3D software. Good lighting is the difference between bad CG and mind-blowingly stunning CG. Computer game designers are only just coming to realise that it's not textures or graphics that make games look real, but the patterns of light and dark. That's why dark games (or dark scenes) look almost real, while bright interiors are obviously fake.

There already exist noise analysing software to ID faked digital images. Cant remember the name though, it would be a great proggy to buy.

reply to post by watch_the_rocks


I recommend checking out the new hit game called Crysis. This is BY FAR the most realistic game EVER!

And your right, it comes down to lighting. Textures can only go so far. Its all about depth.

"I would be nice to use this tool to determine if any of the Apollo moon landing photos were faked. Theoretically there should only be a single light source, shouldn't there?"

Ug... how do these people not accidentally trip and kill themselves just getting out of bed in the morning. Even if the Apollo images and videos were faked, they would have been shot on a set... not edited in Photoshop!

Originally posted by NGC2736
If you're the cops, and you're sure that Joe Shmoe is the guy holding up liquor stores around town, can you fake a store video to use as evidence?

That kind of rendering is still difficult to accomplish even to fool the eye. Gets expensive too. I wouldn't put it past the feds for a minute, but local PD doesn't have the resources. Not for a piddly hold-up perp anyway. Cop killer though... Hell they'd edit in their own grandmother.

Originally posted by apc
"Even if the Apollo images and videos were faked, they would have been shot on a set... not edited in Photoshop!

Heres the thing....

If you had a tool that lets you figure out light sources, wouldn't you be able to tell where lights were placed if used, and how many light sources exist in frame?

It doesn't matter if it wasn't photoshopped, as long as the light source(s) were inconsistent to that of the sun.

If you told me you had something like that I'd assume that you need to specify reflectivity of all objects in sight.

Something like HDR, where people set up a metal ball to catch the lighting in the area.

-0mega-, HDR stands for High Dynamic Range, and is a digital process whereby the photographer takes multiple images of the same scene at different exposure lengths, and combines them using special software.
The metal ball you're referring to is just an example of how this works, as a metal ball is highly reflective and is good for demonstrations.

Originally posted by IMAdamnALIEN
If you had a tool that lets you figure out light sources, wouldn't you be able to tell where lights were placed if used, and how many light sources exist in frame?

Sure. But that doesn't tell you anything as far as image manipulation. Any other theories have been debunked time and time again. Refer to the Moon Hoax thread for that subject.

What you're looking for in manipulated images is different light sources for different image elements. If object A is casting several shadows from multiple light sources but object B is casting different shadows from different light sources, one of the objects may have been edited in. Some of the shadows in the entire image just looking strange because of environmental conditions is not an indicator of manipulation. So no... analyzing one of the Apollo shots wouldn't tell you anything unless a light source for say the flag can not be found casting light on Buzz, suggesting either the flag or Buzz, or both, was not in the original image.

reply to post by watch_the_rocks

Originally posted by watch_the_rocks
This is a fun little test you can take: Fake or Photo?

I got every single one right, excluding one image in the bonus round.

But I've been using photoshop for years, and can spot a fake pretty easily. Yes, the main thing is lighting. It is very very hard to replicate, I would say almost impossible without proper 3D software.

That was fun! The first 10 are easy. Ten years ago I thought I would always be able to spot CGI animations in movies. I changed my mind recently.
I don't think a generic tool can detect most fakes, or even a small percentage with any accuracy. Besides, there are so many ways to create fake images, other than CG.

reply to post by apc


I think you may be missing my point. The moons images DO NOT have to be manipulated. All we want to find out is light sources. Photoshop didn't exist in 1969 so how could it be a manipulated other than physical methods?

Get photographic manipulation out of your mind for the moment, even though the thread is about it..... Use the tool for other purposes......Do you not see this?

I'm not looking for something added in.....Its all about the light....I don't care about the rest.......The lighting has the answer and this tool can help sift it out.

Problem with using light sources to determine if the picture is faked, is the picture is 2 dimensional. So the only way you can know where light is refracting from off of an object, is to know the exact 3 dimensional shape of that object, and it's distance from the lens.

If you already know the actual shape and size of the object in the photograph, then yeah, great. But without that, the objects faces could be on any number of angles, placing the light source in question pretty much anywhere.

reply to post by IMAdamnALIEN

I'm not bothering because like you said, that's not what this thread is about. What you talk about has been beaten to death in the Moon Hoax thread. It's in Space Exploration. Big thread, can't miss it.