Liveness Detection

Real person.Real moment.No exceptions.

Active and passive liveness detection that blocks printed photos, masks, deepfakes, and video replay attacks — certified to iBeta Level 1 & 2. Keeps fraudsters out without adding friction for real users.

Request a demo

Active Liveness · Challenge Running

😊

Please turn your head left slowly

✓ Blink✓ Smile↻ Turn LeftLook Up

3 of 4 challenges complete

Depth map verified

REAL

Texture analysis

LIVE SKIN

Head pose tracking

ACTIVE

Reflection artefacts

NONE

Print attack signal

NOT DETECTED

Liveness Result
Confidence: 99.7% · 840ms

✓ LIVE

99.9%

Liveness accuracy

iBeta L1+L2

Certified

<1s

Passive mode

0.01%

False accept rate

Detection Methods

Two modes. One unbreakable defence.

DigiVerify offers both active and passive liveness detection — choose based on your use case, risk level, and the user experience you want to deliver.

● Active Liveness

Challenge-Response Detection

The user is presented with randomised micro-challenges — blink, smile, turn head, look up — that must be completed in a specific order. Impossible to spoof with a static photo, pre-recorded video, or 3D mask because each session generates a unique challenge sequence.

Randomised challenge sequences — no two sessions identical
Blink, smile, head turn, look up/down challenges
Head pose estimation with 3D landmark tracking
Impossible to replay-attack with recorded video
Best for: high-risk onboarding, transaction auth, account recovery

⚡ Recommended for high-risk flows

● Passive Liveness

Silent AI Analysis

No challenges. The user simply looks at the camera and takes a selfie. DigiVerify's AI analyses the single frame for depth cues, micro-texture patterns, skin reflectance properties, and artefacts introduced by screens or printing — determining liveness invisibly.

Zero user friction — single selfie capture
Depth map estimation from a single 2D image
Skin texture & reflectance analysis
Screen moiré and print artefact detection
Best for: low-friction onboarding, re-auth, low-risk flows

✓ Recommended for frictionless UX

Threats We Block

Every spoof attack. Detected and rejected.

DigiVerify is tested and certified against the full spectrum of presentation attacks — from basic printed photos to sophisticated 3D masks and AI-generated deepfakes.

Printed Photo Attack

A printed photograph of the target user held in front of the camera. The most common, cheapest spoof attempt.

Blocked · Active & Passive

Screen Replay Attack

A photo or video of the target displayed on a phone or tablet screen, held in front of the camera during verification.

Blocked · Active & Passive

3D Mask Attack

A silicone or resin 3D mask of the target's face worn by the attacker. More sophisticated but etectable via depth and texture analysis.

Blocked · Active Mode

Deepfake / AI Face Swap

AI-generated video or real-time face-swap filters applied during a live camera session. Detectable via artefact analysis and challenge randomisation.

Blocked · Active Mode

How It Works

From camera open to liveness verdict — in under a second.

The liveness pipeline analyses every frame of the capture in real time, combining multiple AI signals into a single confidence-scored decision.

Camera Initialisation

The SDK opens the device camera and immediately begins frame analysis. It checks ambient lighting quality, face position, distance from camera, and angle — guiding the user into the optimal capture position before any challenge begins.

Challenge Generation (Active) / Frame Capture (Passive)

Active: a randomised sequence of micro-challenges is generated server-side and delivered to the SDK. Passive: a single high-quality frame is captured silently once the face is correctly positioned.

Multi-signal AI Analysis

Six parallel AI signals are computed: depth map estimation, skin texture analysis, 3D landmark tracking, reflectance properties, artefact detection (moiré patterns, screen edges), and motion compliance for active challenges.

Liveness Decision

A weighted composite score is computed across all signals. You receive LIVE / SPOOF with a confidence score, per-signal breakdown, attack type flag if spoofing is detected, and the best-quality selfie frame for downstream face matching.

Camera Initialisation

😐

Checking environment...

Face detected

Lighting: Adequate

Distance: Optimal

Hold steady...

Challenge Generation

Active Challenge Sequence

Blink twice — completed (0.8s)

Smile — completed (1.1s)

Turn head left — in progress...

Look up — waiting

Session token: ld_8f4k2m · Expires: 120s

Multi-signal AI Analysis

AI Signal Analysis

Depth map99.1%

Skin texture98.4%

3D landmarks99.7%

Reflectance97.2%

ArtefactsNone

Motion comply100%

Liveness Decision

LIVEConfidence: 99.7%

{
  "decision": "LIVE",
  "confidence": 99.7,
  "mode": "ACTIVE",
  "response_ms": 840,
  "attack_detected": false,
  "attack_type": null,
  "signals": {
    "depth_map": 99.1,
    "skin_texture": 98.4,
    "landmarks_3d": 99.7,
    "motion_comply": 100.0
  },
  "best_frame_url": "https://..."
}

CERTIFICATIONS

Independently tested. Industry certified.

DigiVerify's liveness detection is certified under the most rigorous biometric standards in the industry — not self-assessed.

iBeta Level 1

PAD testing per ISO 30107-3 — photos, screens, basic masks

iBeta Level 2

Advanced attack testing — 3D masks, silicone faces, deepfakes

ISO 30107-3

International standard for biometric presentation attack detection

Bias Audited

Equal performance across all demographics and skin tones

99.9% liveness accuracy

Correctly identifies real users across all tested conditions — different lighting, skin tones, ages, and devices.

0.01% false accept rate

Spoof attacks that slip through at the default threshold. Configurable to 0.00% for highest-risk use cases.

0.5% false reject rate

Real users incorrectly rejected. Well below the 1% industry benchmark — minimises legitimate user drop-off.

<1s passive · 4–6s active

Passive liveness returns a decision in under 1 second. Active mode completes the full challenge sequence in 4–6 seconds.

Works on any smartphone

No special hardware required. Runs on standard front cameras — no depth sensor, no IR emitter, no lidar needed.

For Developers

Session-based API.SDK does the heavy lifting.

Your backend creates a liveness session. The SDK handles camera, challenges, and frame capture. You retrieve the result. No video processing in your codebase.

Chained with Face Verification

The best_frame_url from liveness is passed directly as the selfie_image in face/match. One flow: liveness → face verify → done.

Configurable challenges

Choose which challenges to include and how many. Scale friction to match your risk level — 1 challenge for low-risk, 4 for high-risk flows.

Pre-built SDK UI

iOS, Android, and Web SDKs include the full liveness UI — challenge overlay, guidance text, progress indicator, and retry logic

Webhook on completion

Get pushed to your backend the moment a liveness session completes — no polling, no waiting

liveness-check.js

// DigiVerify — Liveness Detection API

// Step 1: create a liveness session
const session = await fetch('https://api.DigiVerify.com/v1/liveness/session', {
  method: 'POST',
  headers: { 'Authorization': `Bearer ${apiKey}` },
  body: JSON.stringify({
    mode: 'ACTIVE', // or 'PASSIVE'
    challenges: ['BLINK', 'SMILE', 'TURN_LEFT']
  })
});

const { session_id, sdk_token } = await session.json();

// Step 2: SDK uses session_id to run challenge
// (handled automatically by iOS/Android/Web SDK)

// Step 3: retrieve result when SDK completes
const result = await fetch(
  `https://api.DigiVerify.com/v1/liveness/${session_id}`,
  { headers: { 'Authorization': `Bearer ${apiKey}` } }
);

const {
  decision,       // 'LIVE' | 'SPOOF'
  confidence,     // 0-100
  attack_detected, // boolean
  best_frame_url  // for face matching
} = await result.json();

Liveness is the first line of defence.Pair it for the full picture.

Liveness alone confirms the person is present. Add face matching and document verification to confirm they are who they claim to be.

Face Verification

Once liveness is confirmed, match the captured selfie against the reference photo on the identity document. Liveness + face match is the gold-standard biometric pair.

Learn more

Document Verification

Authenticate the identity document and extract the biometric photo — then run liveness and face matching to confirm the complete identity chain.

Learn more

AML Screening

With identity biometrically confirmed via liveness and face match, screen the verified person against 1,500+ global watchlists to complete a full KYC flow.

Learn more