EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Deep Fake Video Detection model: Fine-tuned with Mobile NetV2 and CelebDF-v2 Dataset

Milind Rane, Samiksha Solanke, Rutika Shripati, Rupesh Sonawane, Swaraj Patil, Mangesh Tale

Journal of Fuzzy Intelligence Modelling Optimisation · 30.06.2026 · 1 (2) · 1-6

DOI: 10.13074/jfimo.2026.06.262002 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Fuzzy Intelligence Modelling Optimisation, online erschienen am 30.06.2026. Verfasst von Milind Rane und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The rapid advancement of Generative Adversarial Networks (GANs) has accelerated the proliferation of deep fake technologies, posing significant threats to cyber security, privacy, and public trust. While deep fakes are widely used for entertainment and satire, they are increasingly exploited for malicious activities, including financial fraud, identity theft, and political disinformation. The high realism and sophistication of modern deep fake videos make manual detection impractical, particularly in real-time applications such as video conferencing and live streaming. This paper presents an efficient deep learning-based framework for real-time deep fake detection using the MobileNetV2 architecture, fine-tuned on the Celeb-DF v2 dataset. The proposed approach emphasizes lightweight deployment on mobile devices through Tensor Flow Lite, making it suitable for resource-constrained environments. The methodology includes frame extraction, data preprocessing, model training, and optimization for low-computational-cost inference. Experimental results demonstrate that the proposed model achieves over 83% validation accuracy and 81% test accuracy while maintaining low computational overhead. These findings highlight the potential of the proposed framework for enhancing mobile-based cyber security applications by enabling efficient and real-time detection of manipulated media.

Zitieren

APA

Rane, M., Solanke, S., Shripati, R., Sonawane, R., Patil, S., & Tale, M. (2026). Deep Fake Video Detection model: Fine-tuned with Mobile NetV2 and CelebDF-v2 Dataset. Journal of Fuzzy Intelligence Modelling Optimisation, 1(2), 1-6. https://doi.org/10.13074/jfimo.2026.06.262002

Vancouver

Rane M, Solanke S, Shripati R, Sonawane R, Patil S, Tale M. Deep Fake Video Detection model: Fine-tuned with Mobile NetV2 and CelebDF-v2 Dataset. Journal of Fuzzy Intelligence Modelling Optimisation. 2026;1(2):1-6. doi:10.13074/jfimo.2026.06.262002

BibTeX
@article{Rane2026,
  author = {Rane, Milind and Solanke, Samiksha and Shripati, Rutika and Sonawane, Rupesh and Patil, Swaraj and Tale, Mangesh},
  title = {Deep Fake Video Detection model: Fine-tuned with Mobile NetV2 and CelebDF-v2 Dataset},
  journal = {Journal of Fuzzy Intelligence Modelling Optimisation},
  year = {2026},
  volume = {1},
  number = {2},
  pages = {1--6},
  doi = {10.13074/jfimo.2026.06.262002}
}
Kostenlos starten →