Deep Fake Detection using Advance ConvNets2D

Even a few decades ago, it would have been impossible to foresee the widespread use of deep learning to tackle such complex issues. It has many positive applications but it can also be misused to cause harm to the society. One such challenge is deep fakes, and now more than ever, when anyone with a smartphone app can create a fake image or video, it is crucial to take measures to authenticate the content online. Visual and auditory “deepfakes” may be generated with the aid of neural networks. Deepfake mimics machine learning models with the help of Generative Adversarial Networks. Many well-known people have been affected by the rapid spread of false news stories on social media platforms like Facebook, YouTube, Twitter, etc. Artificial Neural Network (ANN) based deep fakes may appear identical to authentic images or videos before being moderated but they always have spatial and temporal signatures. A neural network has been trained to specialize in Deep fake detection can quickly identify these signs. To determine whether or not a video is deep fake, Residual Network and InceptionResNetV2 models are employed in this research.