FACE RECOGNITION-BASED ONLINE VOTING SYSTEM USING DJANGO AND OPENCV

The integrity of electoral processes is a cornerstone of democratic governance. This paper presents a face
recognition-based Online Voting System developed using the Django web framework and OpenCV computer
vision library. The system implements a two-factor authentication mechanism — combining password-based
login with real-time webcam face verification — to ensure voter identity before casting a ballot. Face verification
is performed using Haar Cascade classifiers for face detection and HSV histogram correlation for biometric
matching, eliminating the need for external hardware or deep learning models. A database-enforced one-vote-per�voter constraint prevents duplicate submissions. The system provides a real-time results dashboard with Chart.js
visualisation. Experimental evaluation demonstrates reliable face matching under standard indoor lighting
conditions, with a correlation threshold of 0.55 achieving an acceptable balance between false acceptance and
false rejection. The proposed system offers a lightweight, open-source, and deployable alternative to costly
proprietary e-voting solutions, making secure digital voting accessible to academic institutions, small
organisations, and local governance bodies with limited IT infrastructure