Biometrics Based Secured Online Voting System Using Machine Learning Method

EasyChair Preprint 8173

Abstract

Electronic voting or e-voting has been used in varying forms since the 1970s with fundamental benefits over paper-based systems such as increased efficiency and reduced errors. However, there remain challenges to achieving widespread adoption of such systems, especially with respect to improving their resilience against potential faults. Over the course of time, electronic voting has evolved as a substitute for paper ballot voting to decrease redundancies and inconsistencies. Due to the many security and privacy vulnerabilities experienced over time, the past results of e-voting in the last three decades indicate that it has not been very successful. A novel hybrid design-based electronic voting system is proposed, implemented, and analyzed. The proposed system uses two voter verification techniques to give better results in comparison to single identification-based systems. Fingerprint and facial recognition-based methods are used for voter identification. For fingerprint verification MANTRA MFS 100 device is used and for facial recognition, the automatic voting system uses Convolutional Neural Network (CNN). Also, all votes are encrypted using a homomorphic-based Paillier cryptosystem. This work is targeted to replace the manual verification system with a biometric verification system. The developed system also examines carefully whether the voter has voted once or more.

Keyphrases: Fingerprint Verification, e-voting, face recognition, homomorphic encryption

@booklet{EasyChair:8173,
  author    = {Tazaeen Shaikh and Hritika Ranadhir and Suyash Gugale and Vrushali Patil and Omkaresh Kulkarni},
  title     = {Biometrics Based Secured Online Voting System Using Machine Learning Method},
  howpublished = {EasyChair Preprint 8173},
  year      = {EasyChair, 2022}}