Verisoul Docs


Verisoul uses facial biometrics to help ensure the user is a real, live and unique person.

Process Overview

Face Overview
  1. 1.
    First, Verisoul captures a FaceScan. The FaceScan includes two important types of information
    • Extracted 3D vectors representing a key facial features and landmarks
      • Three dimensional vectors are crucial as it prevents a malicious user from presenting a picture or other 2D representation of someone else
    • Liveness data that represent how the user’s face moved during the scan; capturing a FaceScan requires the user to move their face closer to the camera
      • Liveness data captures distortions of the 3D vectors as the movement occurs
  2. 2.
    Next, the FaceScan is encrypted and validated for liveness
    • The 3D vectorized map and liveness data are actually doubly encrypted. Verisoul’s React SDK encrypts the data prior to SSL/TLS standard encryption
    • Using highly trained and fine tuned machine learning models, Verisoul validates that the user data represents a real, live person.
    • Once the user is verified to be real, Verisoul converts the FaceScan to a FacePrint which contains no liveness or image data and cannot be reversed.
  3. 3.
    Verisoul stores NO image or video data. Only an encrypted FacePrint is stored for comparison.
    • Data science models analyze FacePrint in combination with other off-chain and on-chain data to determine user uniqueness.
    • Note: being a non-unique user does not immediately disqualify you from an ecosystem or community as decisioning lies within a customer application. Verisoul’s purpose is to provide ecosystems with the means to make an accurate and well informed decision.