Electronic Know Your Customer (E-Kyc) System

Abstract

The automation of identity verification using digital documents and facial recognition has become important due to the increasing need for secure and efficient electronic Know Your Customer (eKYC) systems. This project focuses on developing an AI-based eKYC verification system that integrates Optical Character Recognition (OCR) and deep learning-based face verification to automate identity authentication using government-issued ID cards such as Aadhaar and PAN cards. The main objective of the system is to extract textual information from ID cards and verify user identity by comparing the face on the ID card with the uploaded face image. The system uses EasyOCR for text extraction and DeepFace with the FaceNet model for face verification. The uploaded ID card image undergoes preprocessing, contour detection, and text extraction to obtain user details such as name, ID number, date of birth, and gender. Face detection and embedding generation are performed on both the ID card image and the uploaded face image, and the similarity between the two faces is calculated to verify identity. The system also performs duplicate verification by checking existing records in the database before storing new user data. The extracted details, hashed ID number, and face embeddings are stored securely in a MySQL database and JSON storage.The proposed system provides an automated, secure, and efficient identity verification solution that reduces manual verification effort, improves accuracy, and prevents duplicate registrations. The modular architecture allows future enhancements such as live face capture and multi-factor authentication, making the system suitable for real-world eKYC applications in banking, finance, and digital identity verification systems.