Machine Learning-Based Multimodal Biometric Authentication System (Facial and Fingerprint Recognition) - Online Voting Systems

Aug 3, 2026

Upload by Riley

Online voting systems deliver convenience over paper ballots, which suffer from logistical constraints, tampering risks, and the requirement for voters to be physically present. Ensuring that every ballot is cast by a legitimate and unique voter is a core security and authentication challenge for preserving electoral integrity. This study proposes a machine learning authentication module integrating facial and fingerprint recognition into a web-based voting platform. A biometric dataset is used to train models, and results are evaluated with accuracy, precision, recall, and R2 score, achieving 98%, 1.0, 0.8, and 0.78 respectively.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages7

Reading Time18 min

AccessFree to download

Introduction

Related Works

ABUAD Journal of Engineering Research and Development (AJERD) ISSN (online): 2645-2685; ISSN (print): 2756-6811 Volume 8, Issue 1, 122-128 | Machine Learning-Based Multimodal Biometric Authentication System (Facial and Fingerprint Recognition) for Online Voting Systems
Solomon OMOZE, Samuel OMAJI, Glory Nosawaru EDEGBE
Department of Computer Science, Edo State University, Uzairue
[omozesolomon@yahoo.com](omozesolomon@yahoo.com/omaji.samuel@edouniversity.edu.ng/edegbe.glory@edouniversity.edu.ng)[/](omozesolomon@yahoo.com/omaji.samuel@edouniversity.edu.ng/edegbe.glory@edouniversity.edu.ng)[omaji.samuel@edouniversity.edu.ng/edegbe.glory@edouniversity.edu.ng](omozesolomon@yahoo.com/omaji.samuel@edouniversity.edu.ng/edegbe.glory@edouniversity.edu.ng) | | --- | | Corresponding Author: [omaji. samuel@edouniversity.edu.ng](omaji. samuel@edouniversity.edu.ng), +2348061932007
Date Submitted: 12/08/2024
Date Accepted: 08/02/2025
Date Published: 11/02/2025 | Abstract: Online voting systems offer many advantages over traditional voting methods, such as paper ballots, this is because paper ballots face some well-known challenges, ranging from logistics, susceptibility to tampering, and the requirement for voters to be physically present at polling stations. In contrast, online voting systems offer the potential to overcome these challenges by providing a convenient and accessible means for citizens to cast their votes from anywhere. However, online voting systems must address significant security and authentication challenges to ensure that each vote is cast by a legitimate and unique voter, maintaining the integrity of the electoral process. This project proposes the development of a machine learning authentication module that can be integrated into an online voting system using facial recognition and fingerprint recognition to enhance the security of online voting. The system therefore consists of two main components; the machine learning-based authentication component and the web-based voting platform. The authentication component uses machine learning algorithms to accurately and reliably verify the identities of voters based on their biometric data. The web-based platform facilitates voter registration, authentication, and voting processes, ensuring a seamless and secure user experience. These two components were implemented first by obtaining a comprehensive database of user biometric data, training the machine learning module, and implementing a user-friendly web interface using Java Server Pages (JSP) and a MySQL database. The system's performance was evaluated using established metrics, including accuracy, precision, recall, and R2 Score with the following values 98%, 1.0, 0.8 and 0.78 respectively. Keywords: Online Voting, Biometric Authentication, Machine Learning, Facial Recognition, Fingerprint Recognition, JSP, MySQL 1. INTRODUCTION As humans depend more and more on computers and digital tools, there has been a growing interest in developing secure and efficient online voting systems. Traditional voting methods, including paper ballots and electronic voting machines, have several limitations, such as logistical challenges, susceptibility to tampering, and the need for physical presence at polling stations. Online voting systems on the other hand, offer the potential to overcome most of these challenges by providing a convenient and accessible means for citizens to cast their votes from anywhere. However, online voting systems face significant security and authentication challenges. Ensuring that each vote is cast by a legitimate and unique voter is critical to maintaining the integrity of the electoral process. Traditional authentication methods, such as usernames and passwords, are insufficient due to their vulnerability to phishing, hacking, and other forms of cyber-attacks. Consequently, there is a need for more robust authentication mechanisms that can provide higher levels of security and re

FAQs

Why are traditional paper-based voting methods considered challenging?

Paper ballots face logistical difficulties, are susceptible to tampering, and require voters to be physically present at polling stations.

What security problem does the proposed system address?

It targets authentication, ensuring each online vote is cast by a legitimate and unique voter to protect electoral integrity.

How does the system implement the authentication module and evaluate performance?

It uses a biometric database to train machine learning models and implements a web interface with JSP and MySQL, then evaluates using accuracy, precision, recall, and R2 score.

Aug 3, 2026

Upload by Riley