Communications of the ACM
A Growing Problem of 'Deepfake Geography': How AI Falsifies Satellite Images
A University of Washington-led study has identified the practice of location spoofing, in which fake photos are created to look like genuine images of real places.
Credit: Zhao et al
Researchers at the University of Washington (UW), Oregon State University, and Binghamton University used satellite photos of three cities and manipulation of video and audio files to identify new methods of detecting deepfake satellite images.
The team used an artificial intelligence framework that can infer the characteristics of satellite images from an urban area, then produce deepfakes by feeding the characteristics of the learned satellite image properties onto a different base map.
The researchers combined maps and satellite imagery from Tacoma, WA, Seattle, and Beijing to compare features and generate deepfakes of Tacoma, based on the characteristics of the other cities.
UW's Bo Zhao said, "This study aims to encourage more holistic understanding of geographic data and information, so that we can demystify the question of absolute reliability of satellite images or other geospatial data. We also want to develop more future-oriented thinking in order to take countermeasures such as fact-checking when necessary."
From University of Washington
View Full Article
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found