Advancements in Artificial Intelligence has been accelerated by advances in computing power and social media's ever-expanding reach and more fake face image generators have emerged worldwide owing to the growth of Face Image Modification (FIM) tools like Face2Face and Deepfake, which pose a severe threat to public trust. High levels of realism can be achieved in these synthesized videos by utilizing generative machine learning models such as Variational AutoEncoders or Generative Adversarial Networks.
Although there have been significant advancements in the identification of certain FIM, a reliable false face detector is still lacking. Convolutional Neural Network (CNN) tends to learn picture content representations because of the structure's relative stability.
The widespread adoption of deepfake technology presents a pressing concern across diverse sectors, encompassing politics, security, and personal privacy. This paper introduces an innovative temporal-aware approach for automatically detecting deepfake videos. Our method employs a dual-stage neural network architecture, comprising a Convolutional Neural Network (CNN) for extracting features at the frame level, followed by a Recurrent Neural Network (RNN) for temporal analysis. By leveraging the inherent temporal dynamics characteristic of deepfake generation, the RNN discerns subtle manipulations to classify videos accurately. We assess the efficacy of our methodology using a comprehensive dataset comprising deepfake videos sourced from various online platforms. Our findings underscore the robustness and competitive performance of our system, underscoring its effectiveness despite its straightforward architecture.
Article Details
Unique Paper ID: 163386
Publication Volume & Issue: Volume 10, Issue 11
Page(s): 831 - 837
Article Preview & Download
Share This Article
Join our RMS
Conference Alert
NCSEM 2024
National Conference on Sustainable Engineering and Management - 2024