Ai-Driven Fraud Detection In E-Kyc With Integrated Fingerprint

Abstract

This project presents the development of a secure and
intelligent e-KYC system integrated with AIdriven fraud
detection and fingerprint authentication. The backend is
implemented using SQLite for lightweight, local storage,
while the admin portal is developed using Django to provide
a visual, scalable, and interactive dashboard for managing
user data. The system performs OCR-based document
extraction from Aadhaar, PAN, and Passport images using
deep learning models to retrieve identity information. To
enhance identity verification, the platform incorporates
fingerprint authentication using pretrained datasets,
allowing biometric comparison with stored entries. A
machine learning-based fraud detection model is trained on
extracted metadata and biometric scores to identify
suspicious entries based on inconsistencies, duplicate IDs,
incomplete fields, or low fingerprint match scores. The
Django
dashboard supports data visualization, search, and secure
export, offering An integrated overview of successfully
verified identities along with entries flagged as potentially
suspicious. Clean UI elements, such as summary cards,
verification badges, and realtime graphs, are used to
display verification metrics and document statistics. This
intelligent system enhances e-KYC operations by reducing
manual errors and delivering accurate fraud detection
through automated AI pipelines. The Django framework
ensures maintainability and flexibility, while the integrated
machine learning components offer smart automation for
secure identity validation in modern digital environments.