FB pixel

Researchers develop “digital watermarks” to detect deepfakes

Categories Biometric R&D  |  Biometrics News

Researchers from the Tandon School of Engineering at New York University are developing methods to identify when an image has been altered using artificial intelligence and “digital watermarks” to help detect deepfakes.

A prototype imaging pipeline increased the chances of detecting manipulation from roughly 45 percent to over 90 percent in tests, without sacrificing image quality, according to the announcement.

Deepfakes are a growing problem, with Samsung researchers recently developing a way to generate them from less data, and Facebook doubling down on fake media as a protected form of free speech.

In an approach pioneered by Department of Computer Science and Engineering Research Assistant Professor Pawel Korus, a typical photo development pipeline was replaced by a neural network, which embeds carefully crafted artifacts which are highly sensitive to manipulation directly into the image at acquisition.

“Unlike previously used watermarking techniques, these AI-learned artifacts can reveal not only the existence of photo manipulations, but also their character,” Korus says.

The process can be performed in-camera, and survives the distortion of online photo sharing services. The technology is open-source.

Most other attempts to analyze image authenticity rely on the final image, but the NYU researchers reasoned that most photos now use machine learning in image acquisition to normalize elements like lighting and stability.

“We have the opportunity to dramatically change the capabilities of next-generation devices when it comes to image integrity and authentication,” says NYU Tandon Professor of Computer Science Nasir Memon, who co-authored the research paper with Korus. “Imaging pipelines that are optimized for forensics could help restore an element of trust in areas where the line between real and fake can be difficult to draw with confidence.”

The researchers will present their paper on “Content Authentication for Neural Imaging Pipelines: End-to-end Optimization of Photo Provenance in Complex Distribution Channels” at the Conference on Computer Vision and Pattern Recognition in June.

Article Topics

 |   |   | 

Latest Biometrics News


UK gov’t introduces new digital identity verification services bill

The UK government has announced a coming identity verification services bill that will support digital ID products and services from…


EES launch postponed to November, EU’s biometric border app may still not be ready

A smartphone app designed to streamline queues for the EU’s upcoming border checks will not be available in time for…


US government transitions Veteran’s Affairs, IRS to Login.gov or ID.me

The U.S. Department of Veterans Affairs (VA) says it will implement a more streamlined login process for veterans to access…


Data breach raises questions about Fractal ID’s decentralized identity architecture

A data breach at decentralized digital identity verification provider Fractal ID has exposed the ID documents and facial images of…


Physical IDs no longer mandatory in Azerbaijan, where 65% use digital identity

Physical and digital IDs have reached parity in Azerbaijan, where the government has announced that identity information provided through the…


Recfaces argues biometric data privacy rule carries Olympics security risk

RecFaces is calling for facial recognition technology (FRT) to be used as a key security measure to protect critical infrastructure…


Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Most Read This Week

Featured Company

Biometrics Insight, Opinion

Digital ID In-Depth

Biometrics White Papers

Biometrics Events