Document liveness detection is software used to detect when an identity document used for digital onboarding is a presentation attack. Document presentation attacks are analogous to facial presentation attacks, where a printed image, mask, or selfie image displayed on a digital screen is used to spoof biometric authentication.
Examples of document presentation attacks are when a document presented during a digital onboarding is either partially or fully comprised of a reproduction or forgery, either physical or digital. Screen replays, printed copies, and portrait substitutions make up the vast majority of real-world document presentation attacks. A screen replay is when a document is reproduced by displaying it on a digital screen. A printed copy is a reproduction of a document printed from a digital image taken by a camera, copier, or scanner. A portrait substitution is when a fraudulent facial photo is overlaid onto the facial portrait of a genuine document.
While accurate and effective document liveness detection is essential, it also should ideally be easy to implement into legacy systems and easy for consumers to use. Customers should not be impeded by friction in the form of complex instructions, failed image captures, and false positives.