Swap It Like Its Hot: Segmentation-based spoof attacks on eye-tracking images

• DOI: 10.48550/arxiv.2404.13827
• Publication Date: 2024-04-21
• Authors: Anish S. Narkar; Brendan David-John
• Keywords: FOS: Computer and information sciences; Computer Vision and Pattern Recognition (cs.CV); Computer Science - Computer Vision and Pattern Recognition

Video-based eye trackers capture the iris biometric and enable authentication to secure user identity. However, is susceptible spoofing another user's identity through physical or digital manipulation. The current standard identify attacks on eye-tracking sensors uses liveness detection. Liveness detection classifies gaze data as real fake, which sufficient detect presentation attacks. such defenses cannot a attack when image inputs are digitally manipulated swap pattern of person. We propose IrisSwap novel gaze-based allows attackers segment in victim's fool authentication. Both offline online produce that deceives state-of-the-art defense models at rates up 58% motivates need develop more advanced methods for trackers.