Comprehensive Evaluation of Deepfake Detection Models: Accuracy, Generalization, and Resilience to Adversarial Attacks
Applied Sciences · 25.01.2025 · 15 (3) · 1225
DOI: 10.3390/app15031225 ↗Kurz gefasst
Fachartikel in der Zeitschrift Applied Sciences, online erschienen am 25.01.2025. Verfasst von Maryam Abbasi und 3 weiteren Autor:innen. Die Arbeit wurde bisher 35-mal zitiert (Crossref) und nennt 22 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: FCT—Foundation for Science and Technology, I.P., Research Center in Digital Services (CISeD), Instituto Politécnico de Viseu.
Abstract (Originalsprache)
The rise of deepfakes—synthetic media generated using artificial intelligence—threatens digital content authenticity, facilitating misinformation and manipulation. However, deepfakes can also depict real or entirely fictitious individuals, leveraging state-of-the-art techniques such as generative adversarial networks (GANs) and emerging diffusion-based models. Existing detection methods face challenges with generalization across datasets and vulnerability to adversarial attacks. This study focuses on subsets of frames extracted from the DeepFake Detection Challenge (DFDC) and FaceForensics++ videos to evaluate three convolutional neural network architectures—XCeption, ResNet, and VGG16—for deepfake detection. Performance metrics include accuracy, precision, F1-score, AUC-ROC, and Matthews Correlation Coefficient (MCC), combined with an assessment of resilience to adversarial perturbations via the Fast Gradient Sign Method (FGSM). Among the tested models, XCeption achieves the highest accuracy (89.2% on DFDC), strong generalization, and real-time suitability, while VGG16 excels in precision and ResNet provides faster inference. However, all models exhibit reduced performance under adversarial conditions, underscoring the need for enhanced resilience. These findings indicate that robust detection systems must consider advanced generative approaches, adversarial defenses, and cross-dataset adaptation to effectively counter evolving deepfake threats.
Förderung
- FCT—Foundation for Science and Technology, I.P.
- Research Center in Digital Services (CISeD)
- Instituto Politécnico de Viseu
Zitieren
APA
Abbasi, M., Váz, P., Silva, J., & Martins, P. (2025). Comprehensive Evaluation of Deepfake Detection Models: Accuracy, Generalization, and Resilience to Adversarial Attacks. Applied Sciences, 15(3), 1225. https://doi.org/10.3390/app15031225
Vancouver
Abbasi M, Váz P, Silva J, Martins P. Comprehensive Evaluation of Deepfake Detection Models: Accuracy, Generalization, and Resilience to Adversarial Attacks. Applied Sciences. 2025;15(3):1225. doi:10.3390/app15031225
BibTeX
@article{Abbasi2025,
author = {Abbasi, Maryam and Váz, Paulo and Silva, José and Martins, Pedro},
title = {Comprehensive Evaluation of Deepfake Detection Models: Accuracy, Generalization, and Resilience to Adversarial Attacks},
journal = {Applied Sciences},
year = {2025},
volume = {15},
number = {3},
pages = {1225},
doi = {10.3390/app15031225}
}
EUVIMED