The Video Selfie module records a short video of the user performing a guided series of actions, including capturing their ID and selfie, to confirm physical presence and run liveness and face match checks. It can also capture voice consent. It is commonly used by financial institutions to meet regulatory requirements for remote identity verification. The module digitally replicates what a branch manager would do in person.
Integrations
:white_check_mark: Web | :white_check_mark: iOS | :white_check_mark: Android
How It Works
Video Selfie is a multi-check verification step, not simply a video recording. It depends on the ID Capture and Face Capture modules being completed earlier in the session. Those modules provide the reference data for all the comparisons Video Selfie performs.
During Video Selfie, the user records a short video in which they are guided to show their face. Depending on configuration, they may also be guided to show their identity document and provide spoken consent. The recording is then processed against the following checks:
Checks that contribute to the Video Selfie score:
| Check | What it does |
|---|---|
| Speech Recognition | Transcribes the user's spoken consent and validates it against the expected statement. Only runs when Voice Consent is enabled. |
| Face Recognition | Compares the selfie captured during Video Selfie against the selfie from the prior Face Capture module step. You can configure sensitivity using Face Recognition Severity. |
| Liveness | Confirms the user is physically present, not a photo or spoof, in real time during recording. |
Informational checks (do not affect score by default):
| Check | What it does |
|---|---|
| Face Match | Compares the ID photo from ID Capture against the ID shown during Video Selfie. |
| Type of ID Match | Confirms the same document type is used throughout the session. |
| OCR Match | Compares the full name extracted from the ID shown in Video Selfie against the original OCR data. |
| Back ID Type Match | Performs the same document-type check for the back of the ID. |
| Back ID OCR Match | Compares back-of-ID data against original OCR data. |
| Voice Consent Face Recognition | Validates that the person giving voice consent matches the prior selfie. Only runs when the Voice Consent setting is enabled. |
| Hat Check | Detects whether the person is wearing a hat or other head covering during the selfie. |
| Lenses Check | Detects whether the person is wearing glasses or sunglasses during the selfie. |
| Mask Check | Detects whether the person’s face is covered by a mask during the selfie |
| Closed Eyes Check | Detects whether the person’s eyes are closed during the selfie. |
The Video Selfie score is the average of the three scoring checks above. Custom decisions based on any check can be configured to affect either the Video Selfie score or the total session score.
Captured video is uploaded asynchronously after recording. The Incode Platform checks for the file every 20 seconds, up to a configurable maximum window. The default window is 4 minutes. You can configure a webhook to notify your system when processing is complete.
Use Video Selfie
For instructions on configuring your Flow or Workflow to use Video Selfie and then implementing it in your supported integration, use the following pages: