Real or virtual: Scientists ask—can we tell the difference?

February 18, 2016
A Dartmouth College study shows that people find it increasingly difficult to distinguish between computer-generated images and real photos, but that a small amount of training greatly improves their accuracy. Credit: Dartmouth College

A Dartmouth College-led study shows that people find it increasingly difficult to distinguish between computer-generated images and real photos, but that a small amount of training greatly improves their accuracy.

The findings, which have implications for the legality and prosecution of child pornography, appear in the journal ACM Transactions on Applied Perception.

As 3-D rendering software and hardware become more powerful, the computer-generated characters they create for film making, video games, advertising and other venues have become more photo-realistic. But the drive to create virtual characters that are indistinguishable from human characters has also given rise to complex forensic and legal issues, such as the need to distinguish between computer-generated and photographic images of child pornography, says senior author Hany Farid, a professor of computer science and a pioneering researcher in digital forensics at Dartmouth.

"As quickly become more realistic, it becomes increasingly difficult for untrained human observers to make this distinction between the virtual and the real," Farid says. "This can be problematic when a photograph is introduced into a court of law and the jury has to assess its authenticity."

Real or virtual: Dartmouth scientists ask -- can we tell the difference?
A Dartmouth College study shows that people find it increasingly difficult to distinguish between computer-generated images and real photos, but that a small amount of training greatly improves their accuracy. Credit: Dartmouth College

Legal background:

  • In 1996, Congress passed the Child Pornography Prevention Act (CPPA), which made illegal "any visual depiction including any photograph, film, video, picture or computer-generated image that is, or appears to be, of a minor engaging in sexually explicit conduct."
  • In 2002, the U.S. Supreme Court ruled that the CPPA infringed on the First Amendment and classified computer-generated child pornography as protected speech. As a result, defense attorneys need only claim their client's images of child pornography are computer generated.
  • In 2003, Congress passed the PROTECT Act, which classified computer generated child pornography as "obscene," but this law didn't eliminate the so-called "virtual defense" because juries are reluctant to send a defendant to prison for merely possessing computer-generated imagery when no real child was harmed.

In their new study, Farid's team conducted perceptual experiments in which 60 high-quality computer-generated and photographic images of men's and women's faces were shown to 250 observers. Each observer was asked to classify each image as either computer generated or photographic. Observers correctly classified photographic images 92 percent of the time, but correctly classified computer-generated images only 60 percent of the time.

In a follow-up experiment, the researchers found that when a second set of observers was provided some training prior to the experiment, their accuracy on classifying photographic images fell slightly to 85 percent but their accuracy on computer-generated images jumped to 76 percent.

With or without training, observers performed much worse than Farid's team observed five years ago in a study when computer-generated imagery was not as photo-realistic.

"We expect that as computer-graphics technology continues to advance, observers will find it increasingly difficult to distinguish computer-generated from ," Farid says. "While this can be considered a success for the computer-graphics community, it will no doubt lead to complications for the legal and forensic communities. We expect that human will be able to continue to perform this task for a few years to come, but eventually we will have to refine existing techniques and develop new computational methods that can detect fine-grained image details that may not be identifiable by the human visual system."

Explore further: Swedish research can make Super Mario more realistic

More information: Olivia Holmes et al. Assessing and Improving the Identification of Computer-Generated Portraits, ACM Transactions on Applied Perception (2016). DOI: 10.1145/2871714

Related Stories

Swedish research can make Super Mario more realistic

March 6, 2009

Computer games are being developed at an ever more rapid pace, and the technical demands are rising, not least regarding graphics boards.  At Mälardalen University in Sweden, researchers have now found a solution to ...

How human vision perceives transparent layers

August 7, 2015

(Medical Xpress)—The adaptive advantage of the visual perception of transparency is obvious: the urgent need to find water. Thus, human vision is able to perceive two layers at the same retinal location—a transparent ...

Identifying illegal websites in photos

September 1, 2015

European computer scientists have developed a way to "read" web addresses in images that could improve filters for blocking pornographic, gambling and other sites. They provide details in the new issue of the International ...

Recommended for you

Apple issues update after cyber weapon captured

August 26, 2016

Apple iPhone owners on Friday were urged to install a quickly released security update after a sophisticated attack on an Emirati dissident exposed vulnerabilities targeted by cyber arms dealers.

Auto, aerospace industries warm to 3D printing

August 25, 2016

New 3D printing technology unveiled this week sharply increases the size of objects that can be produced, offering new possibilities to remake manufacturing in the auto, aerospace and other major industries.

4 comments

Adjust slider to filter visible comments by rank

Display comments: newest first

ogg_ogg
not rated yet Feb 18, 2016
This paper is behind a paywall, but their previous paper used pictures of people like Morgan Freeman and a bunch of other CELEBRITIES as "control" pictures! Ridiculous! You can't reach meaningful conclusions if the CG images are compared to celebrity images! This seems to be a hold-over from their use of image processing algorithms to discriminate between the two sets (where it's a sure bet that the chip set doesn't watch movies or TV), but is risible if applied to people's perceptions.
RealScience
not rated yet Feb 18, 2016
The first one looks fake - the shine on the cheek is too uniform, as is the white in the beard.
The second one looks real.

But the article doesn't say, so I can't confirm my suspicion...
promile
Feb 19, 2016
This comment has been removed by a moderator.
Captain Stumpy
5 / 5 (3) Feb 19, 2016
my biggest problem is this part
"This can be problematic when a photograph is introduced into a court of law and the jury has to assess its authenticity."
a jury is not capable of this type of assessment - AND they're not typically provided with the necessary equipment to make said assessments, either

in court, it is the ability of the scientists and professionals witnessing on the evidence that make the authentication and the jury decides whom to believe (prosecution or defense in the case of a separate assessment that is contradictory)... they don't "authenticate" ANY evidence

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more

Click here to reset your password.
Sign in to get notified via email when new comments are made.