Smart Voting System Support through Face Recognition

Abstract

This article presents a novel authentication technique for online voting systems using facial recognition of voters. Presently, India follows two types of voting systems: secret ballot paper and Electronic Voting Machines (EVM). However, both methods have their limitations and drawbacks. Online voting is yet to be implemented in India, and the existing voting system lacks adequate safety and security measures. It requires voters to visit multiple polling booths, resulting in long queues and missed voting opportunities. Additionally, the current system allows ineligible voters to fraudulently cast their votes, leading to numerous issues. Therefore, this project proposes a highly effective and secure voting system. Our approach incorporates three levels of security in the voting process. 

The first level involves verifying the Unique ID number (UID), followed by the verification of the Election ID number (EID) at the second level. Finally, the third level utilizes face recognition or face matching. By implementing these security measures, our system significantly enhances the security level for each voter. We improve the user authentication process by incorporating face recognition into the application, which accurately determines whether a user is authorized or not.

 

Introduction

 

In India, there are currently two methods of voting. The first method involves a secret ballot paper, which uses numerous paper sheets. The second method is Electronic Voting Machines (EVM), which have been in use since 2003. However, there is a need to propose a more secure method for online voting compared to the existing system. In this article, we propose the use of face detection and recognition to identify the correct person. Our proposed system incorporates three levels of verification for voters. The first level verifies the Unique ID number, the second level verifies the Election Commission ID or voter card number, and the third level utilizes face recognition to match the captured image with the database of face images provided by the Election Commission. If the captured image matches the respective image in the database, the voter is allowed to cast their vote in the election.

 

The existing voting system relies on ballot machines with symbols representing various political parties. By pressing the button with the symbol of the desired party, the vote is cast. However, this system allows for the possibility of fake votes. Individuals may use fraudulent voting cards to cast their votes, resulting in problems. Moreover, voters have to travel long distances to their constituencies to cast their votes. Hence, there is a need for an effective method to identify fraudulent voters during the voting process. Our proposed system addresses these issues and enables voters to cast their votes online, eliminating the need for physical travel.

 

Proposed System

 

In our article concept, we employ three different security levels:

 

Level 1: Unique ID Number (UID)

– During the voter registration process, the system requests a unique ID from the voter. The entered unique ID is verified against the database provided by the Election Commission.

 

Level 2: Election Commission ID Card Number

– In the second level of verification, the voter must enter the Election Commission ID or voter's ID number. The entered ID number is verified against the database provided by the Election Commission.

 

Level 3: Face Recognition with Respective Election Commission ID Number

– This level utilizes the Eigenface algorithm to verify the facial image of the voters from the database provided by the Election Commission.

 

Eigenface Algorithm

 

The Eigenface algorithm follows an appearance-based approach to face recognition. It captures the variation in a collection of face images and encodes individual faces based on this information. The encoded images are compared with the collection of face images in a holistic manner. The Eigenfaces form a basis set of all images used to construct the covariance matrix. A smaller set of basis images is used to represent the original training images, resulting in dimension reduction. By comparing how faces are represented by the basis set, classification can be achieved. Face images are projected into a feature space called "Face Space," which best encodes the variation among known face images. The face space is defined by the Eigenfaces, which are the eigenvectors of the set of faces.

 

Working of Eigenface Algorithm

 

The working flow of the system using the Eigenface algorithm is as follows:

 

1. Initialization: Acquire the training set and calculate Eigenfaces (using PCA projections) that define the Eigen space.

2. When a new face is encountered, calculate its weight.

3. Determine if the image is a face.

4. If it is a face, classify the weight pattern as known or unknown.

5. If the same unknown face is seen several times, incorporate it into known faces (learning process).

6. Principal Component Analysis: Eigenface follows the Principal Component Analysis approach, where the face space forms a cluster in image space.

 

Experiment and Results

 

For our experiments, we utilized facial images from the ORL database, consisting of 16 persons with 10 views each. The training set contained 16×7 images.

Working Flow of the System

 

The working flow of the system involves the following steps:

 

1. Registration: Every new user in India must register for voting. At the time of registration, the system captures the user's face using a web camera and stores the face sample in the server database for security purposes.

2. During the election, three levels of security are implemented: unique ID verification, voter ID verification, and face recognition.

3. The system verifies the entered unique ID and voter ID to ensure their accuracy.

4. If the unique ID and voter ID are correct, the system captures the voter's image and compares it with the respective image in the database or server.

5. If the captured image matches the image in the database, the voter is allowed to cast their vote.

6. On the voting page, buttons representing the participating political parties are displayed. Voters can cast their votes in the election.

7. Once a voter has cast their vote, their ID is automatically logged out, ensuring that each voter can only cast one vote.

8. During the vote counting process, only authorized users from the Election Commission can log in using a secure ID and password. If both the ID and password are correct, the voting process continues.

Conclusion

 

The existing voting system in India suffers from several defects, such as a lengthy process, time-consuming procedures, lack of security, potential for bogus voting, and inadequate security measures. However, our proposed approach offers a highly secure and useful alternative to the existing system. By incorporating three levels of security, we can easily identify false voters and prevent bogus votes during elections. The facial authentication technique plays a crucial role in identifying fraudulent voters, ensuring the integrity of the electoral process. With our proposed smart voting system, voters can cast their votes from anywhere with internet access. This system requires a one-time investment for the government and reduces the need for manpower and resources. The centralized repository allows for easy accessibility of data and enables data backup. The smart voting system provides real-time, updated results, and the database can be updated annually or before each election to enroll new eligible citizens and remove deceased individuals from the voter list.

 

Hashtag/Keyword/Labels:

Smart Voting System, Face Recognition, Authentication, Online Voting, Security, Voter Identification

 

References/Resources:

1. L.Vetrivendan, Dr.R.Viswanathan, J.AngelinBlessy. "Smart Voting System Support through Face Recognition." Seminarsonly.com.

   Link: https://www.seminarsonly.com/computer%20science/smart-voting-system.php

 

For more such Seminar articles click index – Computer Science Seminar Articles list-2023.

[All images are taken from Google Search or respective reference sites.]

 

…till next post, bye-bye and take care.