Identity Verification
Facial Biometrics
Publicly benchmarked face matching, passive liveness detection, and deepfake defence, engineered for diverse populations and low-spec devices across global markets.
Built for Enterprise Integration
Face analysisSelfie Liveness
Face qualityHigh · 92%
Match score99%
NIST
Evaluated liveness detection
Benchmarked
Independently evaluated face matching
Passive & Active
Liveness, no gestures required
How it works
Face matching and NIST-evaluated passive liveness, built for biometric diversity.
01 · Preparation
1
User captures a selfie
The SDK guides the user to a well-lit selfie using passive liveness capture, no head turns or blink instructions.
2
Liveness detection runs
AI models assess the capture for spoofing attempts including printed photos, video replays, deepfake injections, and 3D masks.
3
Face is matched to ID document
The selfie is compared against the portrait extracted from the ID document using independently benchmarked face recognition.
4
Match score and decision returned
A match score, liveness verdict, estimated age, face quality score, and system decision are returned as structured data.
Why Enterprises Choose IDWise for Facial Biometrics
01
Detect impersonation risk
Face matching against the ID document photo catches identity fraud at onboarding before any account is created.
02
Passive liveness
Liveness detection requires no gestures or complex instructions, higher completion rates without sacrificing security.
03
Deepfake defence
Active defences against virtual-camera injection, deepfake video, presentation attacks, and 3D mask spoofing.
04
Nist-evaluated liveness
Passive liveness (presentation-attack detection) evaluated under NIST's independent testing programme. Face matching is independently benchmarked; results available on request.
Capabilities
Robust, enterprise-grade features built for regulated industries
Face Matching
Compares selfie against ID document portrait with match scores. Configurable threshold per risk tier.
Passive Liveness Detection
Verifies the selfie is a live person, not a photo, video replay, 3D mask, or deepfake, without requiring user gestures.
Deepfake & Injection Defence
Detects virtual-camera injection attacks and deepfake video streams at the SDK level before any image is submitted.
Face Quality Analysis
Assesses image quality, occlusion, estimated age, and number of faces detected, returned alongside the match result.
Feature detail
Everything you need for facial biometrics
Passive liveness assesses frame-level signals to detect spoofing without requiring user interaction, optimized for low-spec cameras.
Independently benchmarked matching tuned for diverse skin tones, ethnicities, and facial features across diverse populations and capture conditions.
Camera injection detection, deepfake analysis, and presentation-attack resistance tested against SOTA attack vectors.
Face quality, occlusion detection, estimated age range, and multi-face detection, all returned in the step result.
Built for Enterprise Integration.
Simple API & SDK
Start with a few lines of code. Full sandbox, webhooks for every event, and docs your engineers will use.
Consistent controls across supported channels
Same fraud models, document coverage, and UX on iOS, Android, Web, React Native, and Flutter.
iOSAndroidWebReact NativeFlutterREST API
</> View documentation
// Start a journey with facial-biometrics const journey = await idwise.startJourney({ flowId: 'your_flow_id', referenceNo: 'USR_000001' }); // Receive result via webhook app.get('journey/v2/get/12345667899', (req) => { const { decision } = req.body; // Passed | Refer | Rejected });
Frequently asked questions
IDWise uses passive liveness detection, no head turns, blinks, or gestures required. This improves completion rates significantly while maintaining strong spoofing protection.
Yes. IDWise's passive liveness detection is evaluated under NIST's independent testing programme, and our face-matching models are independently benchmarked; results are available on request.
IDWise detects virtual-camera injection at the SDK level, preventing deepfake video streams from being submitted. Additional frame-level analysis detects presentation attacks.
The system provides real-time capture guidance to improve image quality. If quality remains below threshold after retries, the step fails with a descriptive reason code.
Talk to an IDWise Specialist about Facial Biometrics
Discuss your markets, regulatory requirements, risk controls and integration architecture with our team, and see how IDWise fits your operating model.