Evident's Liveness Verification is the biometric step in Evident's identity verification flow that (1) matches a face from a video selfie to the photo on an identity document and (2) confirms the person completing verification is physically present.
What is Liveness Verification?
Liveness Verification combines two checks:
- Biometric match: compares the face from a video selfie to the photo from the identity document.
- Liveness check: confirms the face in the video selfie belongs to a live person (not a picture, avatar, deepfake, or mask).
As part of the flow, the user is also asked to provide consent to process biometrics in accordance with applicable laws.
What does the user experience look like?
During verification, the user is asked to provide a live selfie and is shown on-screen guidance to help prepare for capture. The flow actively guides users toward a high-quality image — providing real-time prompts for lighting, positioning, and focus. Better captures improve both pass rates and fraud detection accuracy simultaneously.
The biometric step uses passive liveness detection: the user simply looks at the camera and holds still. There are no head-turn prompts, blink commands, or other required movements. Liveness confirmation happens automatically in the background.
If the live selfie cannot be confidently processed, Evident can automatically retry video selfie capture to improve completion (for example, when lighting is poor or the user is wearing sunglasses). Third parties have up to 3 attempts to submit a selfie.
Fraud and attack defense: The biometric engine detects and rejects deepfakes, AI-generated identity attacks, presentation attacks (photos, videos, masks), and camera injection attempts. Behavioral, device, and camera-level checks run alongside the biometric comparison — providing defense against both traditional spoofs and emerging AI-era threats.
How are results returned?
Results are returned as a Valid / Invalid outcome based on the combined biometric match and liveness check.
- If you want a single decision signal, use
Valid/Invalid. - If the fraud rates or your program requirements require more nuance, you can use the combination of confidence level (
High,Medium,Low,None) and status to make a decision.
How does Evident measure fraud risk (FAR)?
The biometric industry standard for fraud measurement is the False Acceptance Rate (FAR). A false accept in biometrics is an instance of a system incorrectly authorizing a person. FAR measures the likelihood that a biometric system incorrectly accepts an access attempt by an unauthorized user. FAR is typically expressed as the ratio of false acceptances divided by the number of identification attempts.
Evident's FAR reference levels for Liveness Verification:
| Score | Confidence Level | FAR | Industry Requirements |
|---|---|---|---|
| 0-54 | None | ||
| 55-69 | Low | 1/1000 | NIST 800-63B minimum for Biometric Matching |
| 70-84 | Medium | 1/10,000 | FIDO Biometric Requirements |
| 85-100 | High | 1/100,000 |
Platform credentials
Evident's identity verification is powered by a platform recognized as a Leader in the Gartner Magic Quadrant for Identity Verification. Independent benchmarks and certifications include:
- ~98% first-try pass rate — the large majority of users verify successfully on their first attempt
- 99% facial-recognition success rate (NIST testing)
- iBeta PAD Level 3 certified for passive liveness on both iOS and Android — 0% error rate (APCER and BPCER)
- Kantara IAL2 compliant; SOC 2 Type II certified; ISO/IEC 30107-3 PAD compliant
- 99.99% platform reliability, processing billions of identity checks annually