Presentation Attack Detection (PAD)

What is a presentation attack?

A presentation attack is an attempt to fool a biometric sensor. The attacker shows a fake sample to the camera or reader. The standard calls this fake sample a presentation attack instrument (PAI).

• —A printed photo of a face
• —A video of a face on a screen
• —A paper mask or a 3D mask
• —A deepfake face on a display

What does the ISO/IEC 30107-3 standard define?

ISO/IEC 30107-3 is the international standard for presentation attack detection. It gives a common taxonomy and a test method. It sets a way to measure how well a system blocks attacks.

Accredited labs test products against this standard. A lab reports a clear pass or fail result.

What do the PAD levels mean?

PAD tests use levels of attack strength. The levels show how hard the fake samples are to make.

• —Level 1 covers simpler attacks. These use low-cost items, such as a printed photo or a screen replay.
• —Level 2 covers stronger attacks. These use custom items, such as a made-to-order 3D mask.

How does PAD relate to liveness detection?

Liveness detection is the common name for this goal. Presentation attack detection is the formal, standard term. PAD adds a taxonomy and a test method from ISO/IEC 30107-3.

The liveness detection page explains the everyday view of the same idea.

How does SenseCrypt use presentation attack detection?

SenseCrypt is a passwordless identity provider (IdP). The user signs in by face. SenseCrypt runs face matching on-device.

The SenseCrypt liveness detection holds iBeta ISO 30107-3 certification. This certification covers Level 1 and Level 2 presentation-attack detection.