Face Match verifies that the applicant in the selfie matches the holder of the ID document (or the previously verified user for repeat actions). It triggers automatically during the Selfie step right after the Liveness check. If the applicant fails the Liveness check, Face Match is not performed.

How Face Match works

Face Match relies on advanced biometric facial recognition to compare an applicant's live selfie against a reference image. The verification process follows three main steps:

1. Feature extraction. Converts detected faces from both the selfie and the reference image (ID document or stored profile) into numerical face embeddings that encode key facial characteristics.
2. Face Comparison. Compares the two embeddings for overall similarity, adjusting for variations in head angle, image scale, and lighting conditions.
3. Match Result. Generates a match or no-match verdict, which appears on the applicant profile and informs the final identity verification decision.

Get started with Face Match

Face Match is automatically included in the Selfie step (Advanced liveness check type) and requires no separate configuration.

To set up the feature:

1. Configure the WebSDK and MobileSDK integration.
2. Create a verification level and add a Selfie step with the Advanced liveness check type.

To review verification results:

1. Navigate to the Applicants page and select an applicant who completed the Selfie step.
2. In the profile, navigate to the Face match section and review the confidence score.

Utilize duplicate detection

While Face Match performs a 1:1 check against a single reference image, Sumsub also offers duplicate detection across your entire user base.

When enabled, the Face matches tab highlights applicants whose selfies are visually similar to the current applicant, even if their profile details (such as name or date of birth) differ.

This helps detect fraud attempts where a single individual tries to onboard using multiple identities.