US-based software company Adobe on Friday announced that it has created a new technology that uses AI and machine learning to help detect images edited using Adobe Photoshop. In the age of rising fake news and "photoshopped" images on social media, the creator of the image-editing software seems to have come to the rescue. The company aims to increase trust and authenticity in digital media. Forensic tools are said to detect manipulation by closely examining elements like noise distribution, strong edges, lighting, and pixel values.
According to Adobe's Senior Research Scientist Vlad Morariu, efforts to detect image manipulation have been on at the firm as part of the DARPA Media Forensics programme since 2016. His research suggests image manipulation can be detected using a series of methods. "We focused on three common tampering techniques - splicing, where parts of two different images are combined; copy-move, where objects in a photograph are moved or cloned from one place to another; and removal, where an object is removed from a photograph, and filled-in," he says explaining the science behind detection.
Detection is based on the simple fact that a manipulated image leaves behind clues that can be reverse studied to reveal how it was altered. While these changes might not be easily visible to the human eye, detection can be done by observing from the pixel level or using filters to highlight them.
First of all, an RGB stream is used, followed by a noise stream filter. As many photographs and cameras have a unique noise signature, inconsistencies can be easily detected. The blog post suggests that while these might be far from the absolute truth, the techniques open up possibilities for making authenticity accountable in the future.
That's not it. Adobe also wants to use human verification to add an additional layer of security. "It's important to develop technology responsibly, but ultimately these technologies are created in service to society. Consequently, we all share the responsibility to address potential negative impacts of new technologies through changes to our social institutions and conventions."