Privacy Preserving E-Voting System Using Homomorphic Encryption

Abstract

With the rise of digital governance, secure and private electronic voting systems are becoming increasingly important. Traditional e-voting platforms often suffer from vulnerabilities such as vote tampering, lack of transparency, and weak identity verification. To address these challenges, this project introduces a secure online voting system that integrates Paillier homomorphic encryption with deep-learning-based facial recognition technology. Homomorphic encryption enables the aggregation of votes while they remain encrypted, ensuring that individual ballots stay completely confidential throughout the counting process. Facial recognition provides real-time voter authentication, effectively preventing proxy and duplicate voting.Throughout the voting lifecycle, ballot data is encrypted during both transmission and storage, eliminating opportunities for third parties to intercept or manipulate information. An administrator can access results only after the voting period concludes, and the system reveals only the final aggregate tally without disclosing individual choices. The proposed model is implemented using Python, Flask, OpenCV, and SQLite, offering a secure, scalable, and user-friendly solution suitable for small-scale elections in academic institutions and local organizations.The system includes an intuitive web interface, encrypted data storage, and secure communication mechanisms to enhance usability and reliability. By combining biometric authentication with advanced cryptographic techniques, the project delivers a robust, tamper-resistant voting framework that strengthens voter trust and ensures endto-end privacy. Overall, this work demonstrates how modern cryptography and artificial intelligence can significantly improve the security and transparency of digital election systems, making it a strong foundation for future e-voting applications.