A Comparative Study of Facial Recognition Models for KYC Verification: FaceJS, ArcFace and Dlib
Year: 2025
Type: conference-paper
Abstract: - In the modern scenario of digitalisation, safe and uninterrupted customer onboarding is crucial for the banking sector, especially for microfinance institutions which cater to the economically marginalised people. This paper develops a Face Verification System using the FaceJS which aims to ease the Know Your Customer (KYC) processes in a Microfinance Banking System. The proposed system captures a live photo of the user being registered and verifies it against the photo extracted from an uploaded government ID duly issued by the government, such as Aadhaar, PAN, or Passport. The FaceJS model adopts facial detection, feature extraction, and comparison techniques to determine the similarity between the two images, leaving very little room for error and hence ensuring security.With its implementation, face verification reduces account openings through fraud, reduces manual intervention, and increases customer's trust in the system. This verification of the customer by face also fulfils the regulatory requirement regarding the dynamic verification of the customer. The paper describes how FaceJS was integrated with Django-based backend services, and it elaborates on the real-time testing results, showing the potential of the application as a reliable and efficient tool for identity verification in the microfinance banking system.
Authors: Abraham, Sharon Mariam, Paul, Dr. Paulin
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W6949510538
FWCI:
Cites:
Cited by:
Related to: