This Adobe Experimental Feature Could Be A Great Tool Against Deep Fakes

Like many tech companies that encourage internal innovation among its engineers, Adobe is not an exception in that it lets engineers every year come up with projects which they call Sneaks. If a Sneak meets internal requirements and would add value to its range of products and services, it is then integrated into the company’s ecosystem.  Maybe the most fascinating Sneak this time around was an element called About Face, which can identify when a face has been tweaked or photoshopped.

In the age of deep fakes and easy manipulation of photos for whatever reasons people do that, it is increasingly difficult to identify the real stuff from the fake. Well Adobe which now says it wants to stem the tide of fake images online, is also a culprit using some of its tools. Well it looks like they eventually see the danger of such tools in the wrong hands and now want to use About Face to proffer a solution.

About Face uses individual pixels instead of using a whole-face recognition algorithm. In light of this can likewise disclose to you which parts of the picture it believes were controlled, giving a virtual heatmap of the modified areas. It appears to be intended to explicitly catch changes made by Photoshop’s liquify feature by attempting to identify areas that may have been extended and compressed to form the new fake image.

Besides, About Face attempts to fix the changes in that it can undo the manipulations. This tool however limited it is can be a useful tool in this era of fake news and deep fakes that have been identified to have far reaching consequences including politics.