Deepfake Detection Based Smart Voting System

Abstract

In recent years, the rapid growth of Artificial Intelligence (AI), Machine Learning (ML) and Deep Learning technologies has made it possible to create realistic fake images and videos known as Deepfakes. They can be misused to create fake identities and spread misinformation. This creates serious challenges in systems where identity verification is important, such as Online or Digital Voting systems. In many regions, especially in Rural areas, fake voting and impersonation during elections remain common problems due to weak verification methods. This research proposes a Deepfake Detection Based Smart Voting System to improve security and transparency of voting process. In the proposed system, first voter use their Voter ID for login. After successful login, the system activates a camera to capture a live images of voter and captured image analysed using a Deepfake Detection Model to determine whether it is real or fake. If image is real, the system then compares live captured image with voter’s profile picture stored in the database using facial recognition techniques. Only when both images match is voter allowed to cast their vote. The proposed system aims to prevent fake voting, detect deepfake attempts and ensure that only genuine voters participate in elections. By combining deepfake detection and facial verification technologies, the system can help to create a more secure, transparent and reliable digital voting process.

Introduction

1. The text discusses the growing threat of deepfake technology, which uses Artificial Intelligence (AI) and Machine Learning (ML) to create highly realistic fake images, videos, and audio. Although deepfakes have useful applications in entertainment and media, they are increasingly misused for spreading misinformation, identity theft, political manipulation, and fraud. Their rapid spread through social media platforms and open-source tools has raised serious concerns about digital security, privacy, and public trust. As a result, deepfake detection has become an important research area in computer vision and cybersecurity.

   The paper proposes a Deepfake Detection-Based Smart Voting System to improve the security of digital elections. Traditional voting systems face challenges such as voter impersonation and identity fraud, especially with the rise of AI-generated fake media. The proposed system integrates deepfake detection and facial recognition to ensure that only genuine voters can cast votes.

   The system works in multiple stages:

2. User authentication using Voter ID and password.

3. Live image capture using a camera to ensure real-time verification.

4. Deepfake detection to check whether the captured image is real or artificially generated.

5. Facial recognition to match the live image with the voter’s stored profile.

6. Secure voting, vote storage, and result management with duplicate vote prevention and data security.

The system adds multiple security layers to prevent fake voting, impersonation, and manipulation while ensuring transparency and voter anonymity. It aims to improve the reliability, integrity, and trustworthiness of digital voting systems.

Conclusion

In this project, a deepfake detection-based secure voting system was developed to enhance the reliability and security of digital voting. The system integrates facial recognition and deepfake detection techniques to verify the identity of voters before allowing them to access the voting platform. A machine learning model was implemented and evaluated using performance metrics such as accuracy, precision, recall, and F1-score, along with a confusion matrix to analyse classification performance. Experimental testing was also conducted under different scenarios, including authentication with a valid user, attempts using another person\'s image, and deepfake image attacks. The results demonstrated that the system successfully allows legitimate users to access the voting page while effectively blocking unauthorized users and deepfake-based attempts. These findings indicate that the proposed approach can improve security, transparency, and trust in electronic voting systems. In the future, the performance of the deepfake detection model can be further improved by using larger and more diverse datasets, as well as applying advanced deep learning architectures to increase detection accuracy. Additionally, the system can be extended for real-time applications, enabling faster and more reliable identity verification in practical voting environments. Future work may also focus on optimizing the model for real-time processing, scalability, and deployment in large-scale voting systems, making it suitable for use in real-world electoral processes and other security-sensitive applications.

References

[1] Aparna Bagde et al., “Deep fake Detection using Deep Learning”, International Journal of Science, Engineering and Technology, 2023, ISSN: 2348-4098 [2] MD Shohel Rana et al., “Deepfake Detection: A Systematic Literature Review”, IEEE Access, Feb 2022, DOI: 10.1109/ACCESS.2022.3154404 [3] Naveen et al., “Visual Deepfake Detection: Review of Techniques, Tools, Limitations, and Future Prospects”, IEEE Access, Dec 2024, DOI: 10.1109/ACCESS.2024.3523288 [4] Peter Edwards et al,. “A Review of Deepfake Techniques: Architecture, Detection, and Datasets”, IEEE Access, Octo 2024, DOI: 10.1109/ACCESS.2024.3477257 [5] Raghava M S et al., “AI DEEP FAKE DETECTION RESEARCH PAPER”, IJNRD, Octo 2023, ISSN: 2456-4184 [6] Rami Mubarak et al., “A Survey on the Detection and Impacts of Deepfakes in Visual, Audio, and Textual Formats”, IEEE Access, Dec 2023, DOI: 10.1109/ACCESS.2023.3344653 [7] Sangho Son et al., “Advancing Generalization in Deepfake Detection: Supervised Contrastive Representation Learning With Dual Stream Spatio-Temporal Features”, IEEE Access, Aug 2025, DOI: 10.1109/ACCESS.2025.3598782 [8] Tackhyun Jung et al., “DeepVision: Deepfakes Detection Using Human Eye Blinking Patter”, IEEE Access, April 2020, DOI: 10.1109/ACCESS.2020.2988660 [9] Yogesh Patel et al., “Deepfake Generation and Detection: Case Study and Challenges”, IEEE Access, Dec 2023, DOI: 10.1109/ACCESS.2023.3342107

Copyright

Copyright © 2026 Shravani Sachin Nalawade, Riya Sandeep Jadhav, Vaishnavi Krishnat Mohite, Tanishka Vaibhav Matane, Mrs. P. D. Ghadge. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.