Glossary · Finance & banking
Liveness check and face match meaning
A liveness check confirms a real person is present; face match confirms it is the person on the document. Learn how both work in Indian KYC, with an example.
How does a liveness check work?
There are two broad approaches. Active liveness asks the person to do something, such as blink, turn their head or read out a number. Passive liveness analyses a single image or short clip for signs of a spoof, such as screen edges, paper texture, reflections or unnatural depth, without asking the person to act. Many systems combine both and return a score or a pass, fail or retry result.
The RBI Master Direction on KYC requires a V-CIP application to include face liveness or spoof detection, and the official on the call must be able to carry out a liveness check. Its digital KYC process, used by an authorised officer in the field, also requires a live photograph of the customer to be captured in the app.
How does face match work?
Face match converts the live face and the document photo into numerical representations and measures how similar they are. The result is a similarity score, not a yes or no. The lender's policy decides the bands: above one level the match is accepted, below another it is rejected, and the range in between goes to a person.
What does this look like on three applicants?
Illustrative example with made-up policy bands: accept at 80 or above, review between 60 and 79, reject below 60.
| Applicant | Liveness | Face match score | Outcome |
|---|---|---|---|
| A | Pass | 91 | Accepted |
| B | Pass | 68 | Sent to a person; the passport photo is ten years old |
| C | Fail | 95 | Rejected; a high match on a photo of a photo is still a spoof |
What is the difference between liveness and face match?
Liveness answers "is a real person here right now". Face match answers "is this the same person as on the document". Passing one says nothing about the other, which is why applicant C above is rejected despite a high match score.
What mistakes do teams make with liveness and face match?
- Using face match alone. Without liveness, a printed photo of the document holder can pass.
- Setting one hard threshold. A single cut-off either rejects genuine customers with old photos or lets weak matches through. A review band handles this.
- Not logging the result. The score, the decision and who overrode it belong in the audit trail.
- Poor capture guidance. Bad lighting and glare cause most retries; tell the customer what to fix.
Where do liveness and face match sit in onboarding?
They are part of the document and identity layer, alongside reading the document and matching its fields. In KYC automation software for banks and NBFCs, liveness and face match sit in that document layer, the lender's own rules and bands decide the outcome, and anything outside the band goes to a person with maker-checker. The maker-checker exceptions playbook covers the review band, and why a KYC agent should start read-only covers controls.
Confirm current liveness requirements for your process in the RBI Master Direction or with your compliance adviser.
Related terms
Read next
Where this shows up in our work. AutoKYC: Documents read on WhatsApp, checked against your policy, exceptions to a person, maker-checker before the core. Book a 30-minute working session with an engineer.
Questions
What is a liveness check in KYC?
A liveness check confirms that a real person is present in front of the camera during KYC, rather than a photo, screen replay, video or mask. RBI's V-CIP standards require liveness or spoof detection.
What is a good face match score?
There is no universal score. Each vendor's scale differs, and the lender's policy sets accept, review and reject bands based on testing with its own customers and document types.
Why did my liveness check fail?
Common causes are poor lighting, glare on glasses, a face partly out of frame, a weak network during a video step, or moving too fast. Retrying in good light usually helps.