Study on Deepfake Face Detection Using Transfer Learning Approach
Jannatul Mawa, Md. Humayun Kabir · International Journal of Computer Applications · 2024
The emergence of deepfake technology has added a new dimension to digital media manipulation.The increasing prevalence of these manipulated visual contents poses significant threats to the authenticity and trustworthiness of digital media.In response to this growing threat, this research work investigates into an approach for detecting deepfake face images through the fusion of transfer learning and deep ensemble neural network techniques.This methodology adopts transfer learning and ensemble neural network techniques to improve the accuracy of detection and generalization capabilities of deepfake detection models.The research includes an extensive evaluation of the deep ensemble neural network on available challenging deepfake datasets.The effectiveness of the applied strategy is evaluated against currently existing techniques using performance indicators such as accuracy, precision, recall, and F1-scores.Finally, this paper presents a notable contribution to the area of deepfake detection through the development of a transfer learning-based deep ensemble neural network.