Home / Current Issue / Paper 1717037
AI-Based Deepfake Detection System: Using Convolutional Neural Networks and Transformer-Based Models
Subject area: Science,Engineering and Technology · Area of research: Convolutional Neural Networks and Transformer
DOI: https://doi.org/10.64388/IREV9I10-1717037
Abstract
The rapid advancement of generative artificial intelligence has led to the widespread creation and distribution of deepfake media — synthetically generated images, videos, and audio that can convincingly imitate real individuals. Deepfakes pose a critical threat to digital trust, misinformation prevention, and public security. Traditional methods of media authentication are no longer sufficient to detect these highly realistic forgeries. This research proposes an AI-based deepfake detection system that integrates Convolutional Neural Networks (CNNs) for spatial feature extraction with a Transformer-based attention mechanism for temporal and contextual analysis. The proposed model is trained on benchmark datasets including FaceForensics++ and DFDC (Deepfake Detection Challenge), enabling it to recognize subtle visual artifacts, inconsistencies in facial geometry, and unnatural blending patterns. Experimental results demonstrate a detection accuracy of 97.3% across diverse manipulation techniques, outperforming existing baseline models significantly. The system also incorporates an explainability module using Grad-CAM visualizations to highlight the regions that contributed to each detection decision. The findings of this study confirm that a hybrid deep learning approach provides a robust and generalizable solution for real-world deepfake detection, contributing meaningfully to the fields of digital forensics, AI ethics, and cybersecurity.
Keywords
Deepfake Detection, Convolutional Neural Network, Transformer Model, Facial Forgery, Digital Forensics, Media Authentication, Grad-CAM, FaceForensics++, Deep Learning
References
[1] T. Tolosana, R. Vera-Rodriguez, J. Fierrez, A. Morales, and J. Ortega-Garcia, “Deepfakes and Beyond: A Survey of Face Manipulation and Fake Detection,” Information Fusion, vol. 64, pp. 131–148, 2020.
[2] A. Rossler, D. Cozzolino, L. Verdoliva, C. Riess, J. Thies, and M. Niessner, “FaceForensics++: Learning to Detect Manipulated Facial Images,” in Proc. IEEE/CVF ICCV, pp. 1–11, 2019.
[3] B. Dolhansky, J. Bitton, B. Pflaum, J. Lu, R. Howes, M. Wang, and C. C. Ferrer, “The Deepfake Detection Challenge (DFDC) Dataset,” arXiv preprint arXiv:2006.07397, 2020.
[4] Y. Li, X. Yang, P. Sun, H. Qi, and S. Lyu, “Celeb-DF: A Large-Scale Challenging Dataset for DeepFake Video Detection,” in Proc. IEEE/CVF CVPR, pp. 3207–3216, 2020.
[5] L. Zheng, J. Shi, J. Zhang, and J. Wu, “Pure Transformers are Powerful Graph Learners for Deepfake Detection,” in Proc. NeurIPS, 2022.
[6] R. R. Selvaraju, M. Cogswell, A. Das, R. Vedantam, D. Parikh, and D. Batra, “Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization,” in Proc. IEEE ICCV, pp. 618–626, 2017.
How to cite this paper
@article{1717037,
author = {Tirth Gajjar, Yashpalsinh Zala},
title = {AI-Based Deepfake Detection System: Using Convolutional Neural Networks and Transformer-Based Models},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {10},
pages = {3679-3683},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1717037.pdf},
abstract = {The rapid advancement of generative artificial intelligence has led to the widespread creation and distribution of deepfake media — synthetically generated images, videos, and audio that can convincingly imitate real individuals. Deepfakes pose a critical threat to digital trust, misinformation prevention, and public security. Traditional methods of media authentication are no longer sufficient to detect these highly realistic forgeries. This research proposes an AI-based deepfake detection system that integrates Convolutional Neural Networks (CNNs) for spatial feature extraction with a Transformer-based attention mechanism for temporal and contextual analysis. The proposed model is trained on benchmark datasets including FaceForensics++ and DFDC (Deepfake Detection Challenge), enabling it to recognize subtle visual artifacts, inconsistencies in facial geometry, and unnatural blending patterns. Experimental results demonstrate a detection accuracy of 97.3% across diverse manipulation techniques, outperforming existing baseline models significantly. The system also incorporates an explainability module using Grad-CAM visualizations to highlight the regions that contributed to each detection decision. The findings of this study confirm that a hybrid deep learning approach provides a robust and generalizable solution for real-world deepfake detection, contributing meaningfully to the fields of digital forensics, AI ethics, and cybersecurity.},
keywords = {Deepfake Detection, Convolutional Neural Network, Transformer Model, Facial Forgery, Digital Forensics, Media Authentication, Grad-CAM, FaceForensics++, Deep Learning},
month = {April},
doi = {https://doi.org/10.64388/IREV9I10-1717037}
}