Inception V3 and GRU Fusion: A Deep Learning Approach for Detecting Deepfake Videos
Vaishnavi Vaishnavi, Ananya Trivedi, Pranjal, Kanika Singla · 2025
As there is a huge rise in the prevalence of Deepfake videos continuously from last few years due to the power being leveraged by some powerful deep learning techniques. Therefore, there is a huge promptness in the global community to develop certain effective methods for Deepfake video detection methods. In this work, the combination of Inception V3 and Gated Recurrent Units (GRU) for spatial feature extraction and temporal dependencies across video frames respectively. Several performance parameters has been used for the quantitative analysis of the models. The result shows that there is notable increase in the accuracy and other parameters which can actually provide a solution to the problem of Deepfake detection.