Skip to content

Appendix A1 — Key Terms

Purpose

This glossary explains the most important face liveness terms in plain English.

Term Simple meaning
Face liveness A check that tries to confirm a real person is physically present during capture
PAD Presentation Attack Detection; the general term for detecting spoof presentation attempts
Bona fide presentation A genuine, legitimate presentation from a real live user
Presentation attack An attempt to fool the system with a fake presentation such as a photo, replay, mask, or injected stream
Passive liveness Liveness detection without asking the user to perform a challenge
Active liveness Liveness detection that asks the user to do something, such as blink or turn
Hybrid liveness A combination of passive and active checks
Threshold A decision cutoff used to turn a score into pass / retry / fail logic
Score fusion Combining multiple model or signal outputs into one decision layer
Retry band / uncertain band Score range where the system does not fully trust the result and asks for retry or escalation
APCER Attack Presentation Classification Error Rate; how often attack samples are wrongly accepted
BPCER Bona Fide Presentation Classification Error Rate; how often genuine users are wrongly rejected
ACER Average of APCER and BPCER; a summary metric often used in PAD discussions
Injection attack A digital bypass where fake media is inserted into the pipeline instead of being physically shown to the camera
Replay attack Showing a previously recorded image or video of the target person on another screen
Challenge-response An active liveness method that asks the user to perform a live action
Manual review A human review path used when automated confidence is not strong enough
Drift A change over time in input conditions or score behavior that affects production performance

Note

Definitions here are simplified for readability. For deeper standards-aligned terminology, see:

Go to Appendix Attack Taxonomy.