Deepfake Detection Using Deep Learning: ResNext and LSTM
Mehzabin Pathan, Smita Khairnar, Sushant Joshi, Hrutvik Malshikare, Avikshit Kharkar, Dattatraya Suryawanshi · 2024
Deep fake technology has been advancing at a rapid pace, and it has become increasingly difficult to distinguish between real and fake videos. Deep fake technology can be used to manipulate and create convincing fake videos that can spread misinformation, propaganda, and even be used for blackmail. As a result, there is a growing need for reliable and accurate deep fake detection techniques to combat this threat. The solution proposed is a deep learning-based approach for deep fake detection. Deep learning models can learn to identify patterns and features in videos that can be used to distinguish between real and fake videos. The proposed system is highlighting the integration of ResNext 50 and LSTM. By exploiting the potential of pretrained models to acquire hierarchical features and combining them with sequential feature extractor such as LSTM detection systems can acquire great precision in differentiating manipulated content from real world. These different ways together contribute to the ongoing work in detecting deep fakes, underlining the need of adopting varied methodologies to successfully prevent the propagation of modified synthetic media.