Skip to main content
Version: 1.18.0 (latest)

Introduction

Biometric Anti-Fraud (BAF) is a biometric identity verification platform with fraud protection, designed for integration into corporate and government remote customer services.

The platform uses artificial intelligence and machine learning for facial recognition, liveness detection, facial image quality assessment, and risk analysis to detect identity fraud attempts.

BAF enables cross-platform user registration and authorization scenarios on web, iOS, and Android using facial biometrics.

BAF versions​

  • Biometric Anti-Fraud (BAF) supports an extended set of scenarios and anti-fraud checks, including user face identification, facial database maintenance, and analysis of additional risk factors.
  • Biometric Anti-Fraud Lite (BAF Lite) is designed for simpler biometric authorization scenarios using 1:1 verification. This version does not support face database maintenance or 1:N search, and also has a limited set of administrative functions.

Example of integrating BAF Lite with an integrator's system​

It is assumed that the BAF Lite web component is initialized by the integrator after the user passes an authentication factor controlled by the integrator.

After passing this factor, the integrator generates a short-lived session token, which is used to initialize the BAF Lite web component.

Integration diagram:

img.png

Types of checks​

Depending on the BAF version, the selected subsystem set, and the usage scenario, the following checks are available:

Basic BAF checks:

  • Comprehensive selfie liveness check
  • Comprehensive facial image quality check
  • Comprehensive check for changes in the user's environment
  • Check against blacklists
  • Check for incidents that trigger risks

Basic BAF Lite checks

  • Comprehensive selfie liveness check
  • Comprehensive facial image quality check

Additional checks for BAF and BAF Lite

  • Checks for signs of DeepFake technology (Experimental functionality)
  • Motion control video saving