Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Learn about a novel security solution presented at USENIX Security '25 that addresses recapture attacks against provenance-based media authentication systems. Discover how sophisticated deepfakes and photo manipulation techniques threaten societal perception of reality, and explore why traditional deepfake detection approaches create an endless arms race between attackers and defenders. Understand the concept of provenance assertion as an alternative approach that combines hardware-based secure camera design with cryptographic authentication to verify the source and processing history of visual content. Examine the critical vulnerability of recapture attacks, where adversaries display fake content on screens or surfaces and use secure cameras to capture seemingly authentic photos and videos. Explore Scoop, a systematic mitigation solution that leverages advanced depth sensing technologies and machine learning-based depth estimation to detect misleading recaptures where display mediums aren't visually identifiable. Review the implementation details on iOS and Android platforms using built-in depth sensors from Apple iPhone 14 Pro and Samsung Galaxy S20 Plus devices. Analyze evaluation results from a comprehensive dataset of 78 recapture attack scenarios, demonstrating Scoop's effectiveness with approximately 95% accuracy on iPhone and 74% accuracy on Samsung devices.
