Security Implementation Protects Users from Secure Authentication Breaches by Hackers
This eye scan security feature removes iris signatures from the digital images used in eye-tracking devices. Eye-tracking functionality will enable many features in the next generation of automobiles, educational hardware, and more. Mixed and virtual reality headsets already use integrated eye trackers in which cameras image the user's eye to detect gaze location and pupil diameter. While this functionality is intended to improve the quality of the user’s experience, built-in eye trackers pose a security threat. Hackers would gain access to a high-resolution image of the user's iris that could potentially breach secure authentication in applications such as banking and voting.
Researchers at the University of Florida have developed a low-cost security step that degrades iris authentication while still allowing accurate gaze tracking in mixed reality headsets and other devices using eye-tracking. The security feature can either defocus the iris in the imaging process or add randomized visual noise to the iris image. Both implementations remove the high frequency identifying features of the iris while preserving the low-frequency components necessary for gaze estimation in applications from gaming to internet security.
Application
Security feature for eye-tracking devices that protects users from spoofing attacks or identity theft by removing iris detail in eye scan images
Advantages
- Adds pixel noise to iris images, preventing hackers from stealing valuable biometric signatures
- Removes only high frequency identifying features in the iris image, preserving low-frequency features to maintain gaze estimation functionality
- Implements as a hardware or software modification, improving the security of eye-tracking devices
- Randomizes visual noise on iris images, providing encryption for higher security applications that use iris recognition such as security firms
Technology
This eye scan security improvement introduces optical defocus to the iris image. Doing so retains the pupil tracking ability of an eye-tracking device while removing high frequency identifying biometric details. An image processing module in a device’s software or a modification of its hardware implements the iris defocusing. Likewise, introducing pixel noise to the image removes a user’s biometric signature. An algorithm randomly selects a percentage of pixels in the iris image and sets them to a constant value, digitally degrading the image. Adding this random noise to the image adds a layer of encryption for contexts requiring a greater level of security.
