There’s no denying that the Internet changed the game. Today, simply by going online, the average person has abilities and privileges that were unfathomable to human beings of past generations.
But that’s not to say those gifts come for free. There are costs involved, and as of late, one of them has been in the spotlight and subject to debate over its potentially harmful ramifications: uncertainty.
At a time when the concept of “fake news” has gone mainstream — and in the midst of a digital landscape in which seemingly anything can go viral — it’s often difficult to genuinely know what’s real.
Fortunately, there are tools available to help. One such tool is FotoForensics, a website dedicated to aiding users in determining the legitimacy of images across a wide array of file formats.
“It works like a microscope — by highlighting artifacts and details that the human eye may not be able to identify,” the site says in describing its service. To use it, all one needs to do is submit an image via upload or link.
The project is the brainchild of computer scientist Neal Krawetz. FotoForensics uses Error Level Analysis (ELA), a method of scrutiny that “permits identifying areas within an image that are at different compression levels,” the site states.
Essentially, ELA can identify the differences in edge, surface, and texture features contained in an image — features that can escape the naked eye. In addition, FotoForensics examines an image’s metadata and can tell you if the file has been previously edited or if it came directly from a camera.
The site even offers tutorials and practice challenges to help users better understand how the FotoForensics system works.
For instance, one challenge highlights the feature discrepancies in several versions of an image of North Korean leader Kim Jong-un that went viral. Another proves that a picture purportedly taken inside Ukrainian president Viktor Yanukavych’s palace directly following his ouster in 2014 was actually shot over a year before.
Still, FotoForensics shouldn’t be viewed as the definitive guide when deciding on image legitimacy, as Krawetz himself notes:
“There is no one-button solution that will tell you if a picture is real or digitally altered. Because situations, content, forums, and questions are as numerous as options to modify pictures, there is no automated analysis result.”
The computer scientist points out that “media outlets like the Associated Press, Reuters, and Getty Images all have different definitions for acceptable alterations.”
Instead, Krawetz says, the site should be used as a launching pad. In terms of validity, the true determining factor resides in each individual’s discernment:
“Without a consistent definition for real, nobody can create a one-button solution. FotoForensics provides tools that help identify alterations. However, it is up to you to determine whether any identified alterations are acceptable for a given picture or field.”
The Homepage of Independent Media