Researchers at the University of California, Riverside (UCR) have developed a deep neural network architecture that can identify manipulated images at the pixel level with high precision.
The researchers labeled nonmanipulated images and the relevant pixels in boundary regions of manipulated images in a large dataset of photos.
The team trained the network on general knowledge about the manipulated and original regions of photos. Then, they tested the neural network on a set of images it had never seen before, and found it was able to detect altered images most of the time.
The researchers say their methodology could be adapted to detect deepfake videos, though there are challenges to overcome.
Said UCR's Amit Roy-Chowdhury, “It's a challenging problem. This is kind of a cat and mouse game. This whole area of cybersecurity is in some ways trying to find better defense mechanisms, but then the attacker also finds better mechanisms."
From University of California, Riverside
View Full Article
Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA