PILLAR 02 / 04RESEARCH DIRECTION

Face & Liveness

Open technologies to protect biometric identity systems against spoofing and emerging attacks.

Read the research questions ↓Collaborate

01 — The basics

Verification and liveness answer different questions

Face verificationIs this the same person as the enrolled record?
Liveness detectionIs a real person physically present in front of the camera?

Both checks are needed. A face match alone can be fooled by a photo, a mask, a screen replay or video injected into the capture pipeline.

02 — Research questions

What we are asking

Q1What separates face verification from proving a live person is present?
Q2How well do defences hold against deepfakes, synthetic identities and injection attacks?
Q3What works reliably on low-cost devices and in low-resource environments?

03 — Areas of work

Where the work is focused

01
Verification vs liveness
Treating matching and presence as separate, testable checks.
02
Presentation attacks
Photos, screens and masks shown to the camera.
03
Deepfakes
AI-generated or altered face video.
04
Synthetic identities
Fabricated identities that combine real and invented data.
05
Injection attacks
Fake video fed directly into the capture pipeline, bypassing the camera.
06
Low-resource settings
Low-cost devices, poor lighting and limited connectivity.

04 — Whitepaper

Open presentation attack detection for identity systems

The case for open, testable liveness and PAD technology, with an evaluation method that works on low-cost devices and in low-resource settings.

STATUS: Draft for comment
VERSION: v0.3 · Sample
LENGTH: ~40 pages

01Verification versus liveness
02The attack landscape
03Open models and benchmarks
04Testing in low-resource settings
05Guidance for procurement

Download draft PDF ↓Comment on the draft

05 — Technology roadmap

From research to open technology

CompleteIn progressPlanned

Q4 2026
Attack dataset collection
Presentation, deepfake and injection samples collected with consent.COMPLETE
Q1 2027
Baseline PAD models
Open models trained and tested on low-cost devices.IN PROGRESS
Q2 2027
Injection attack detection
Signals that separate injected video from real capture.PLANNED
Q4 2027
Public benchmark release
Dataset, protocol and leaderboard for independent testing.PLANNED
Q2 2028
Open PAD toolkit v1.0
Models, tests and integration guides under an open licence.PLANNED

END GOALOpen models, benchmarks and testing methods for presentation attack detection.SAMPLE DATES

06 — Demos

See the technology working

07 — Playbooks

Practical guides for putting it to use

Step-by-step guidance for governments, service providers and builders adopting this pillar’s work.

FOR PROCUREMENT TEAMSSAMPLE
Evaluating a liveness vendor
Questions to ask and tests to require before buying.
Download ↓ Evaluating a liveness vendor20 pages · PDF
FOR TESTING LABSSAMPLE
Running PAD tests on low-cost devices
Set-up, capture conditions and scoring.
Download ↓ Running PAD tests on low-cost devices32 pages · PDF
FOR SYSTEM INTEGRATORSSAMPLE
Handling injection attacks
Hardening the capture pipeline end to end.
Download ↓ Handling injection attacks16 pages · PDF

08 — Papers

Publications from this pillar (4)

Sort

2026
Conference paperSAMPLE
A. Author, B. Author, C. AuthorVenue to be added · 2026
PDF ↓ Presentation attack detection on low-cost mobile devices DOI ↗ Presentation attack detection on low-cost mobile devices
Journal articleSAMPLE
A. Author, B. Author, C. AuthorVenue to be added · 2026
PDF ↓ Detecting video injection attacks in remote identity verification DOI ↗ Detecting video injection attacks in remote identity verification
2025
PreprintSAMPLE
A. Author, B. Author, C. AuthorVenue to be added · 2025
PDF ↓ An open benchmark for deepfake resistance in face verification DOI ↗ An open benchmark for deepfake resistance in face verification
Technical reportSAMPLE
A. Author, B. Author, C. AuthorVenue to be added · 2025
PDF ↓ Verification versus liveness: an evaluation methodology DOI ↗ Verification versus liveness: an evaluation methodology

Entries marked SAMPLE show the layout only. Real publications will be added through the CMS.

Collaborate

Working on biometric security? Talk to us.

dpi@ait.asia

Face & Liveness

Verification and liveness answer different questions

What we are asking

Where the work is focused

Verification vs liveness

Presentation attacks

Deepfakes

Synthetic identities

Injection attacks

Low-resource settings

Open presentation attack detection for identity systems

CONTENTS

From research to open technology

Attack dataset collection

Baseline PAD models

Injection attack detection

Public benchmark release

Open PAD toolkit v1.0

See the technology working

Liveness check in the browser

Deepfake versus genuine capture

PAD benchmark explorer

Practical guides for putting it to use

Evaluating a liveness vendor

Running PAD tests on low-cost devices

Handling injection attacks

Publications from this pillar (4)

Presentation attack detection on low-cost mobile devices

Detecting video injection attacks in remote identity verification

An open benchmark for deepfake resistance in face verification

Verification versus liveness: an evaluation methodology

Get involved in Face & Liveness

Working on biometric security? Talk to us.

Face & Liveness

Verification and liveness answer different questions

What we are asking

Where the work is focused

Verification vs liveness

Presentation attacks

Deepfakes

Synthetic identities

Injection attacks

Low-resource settings

Open presentation attack detection for identity systems

CONTENTS

From research to open technology

Attack dataset collection

Baseline PAD models

Injection attack detection

Public benchmark release

Open PAD toolkit v1.0

See the technology working

Liveness check in the browser

Deepfake versus genuine capture

PAD benchmark explorer

Practical guides for putting it to use

Evaluating a liveness vendor

Running PAD tests on low-cost devices

Handling injection attacks

Publications from this pillar (4)

Presentation attack detection on low-cost mobile devices

Detecting video injection attacks in remote identity verification

An open benchmark for deepfake resistance in face verification

Verification versus liveness: an evaluation methodology

Get involved in Face & Liveness

Working on biometric security? Talk to us.