Secure E-Voting System using Deep Learning Techniques

Democracy is vital in many emerging economies because it places power in control of the general public, who choose government via the electoral procedure. Elections rely heavily on polling places and people. Numerous changes have happened in electoral processes since the year of the republic till now. Subsequently, Ballot sheets were implemented as a voting medium, but this was later replaced by electronic voting machines, in which voters must select from a list of buttons, each symbolizing a symbol. This paper proposes techniques for ensuring voter identification and privacy through fingerprint scanners and other biometric identification techniques-based machine learning algorithms. As a result, we offered a new approach to speed up the voting process and eliminate similar issues. For face identification, we employed Convolutional Neural Network (CNN) architectures such as Alex Net, Visual Geometry Group (VGG-16), and Random Forest Algorithms in this study. The accuracy provided by Alex Net, VGG-16, and Random Forest (R.F.) Algorithms was 94.16%, 98.45%, and 92%, respectively. VGG-16 architecture outperformed all other algorithms with a 98.45% accuracy. We utilized an ATmega328-based Arduino for finger detection.