DEEPFAKE DETECTION USING GAN
A solution based on a GAN discriminator to detect DeepFake videos, which can be used to manipulate public opinion during elections, commit fraud, and discredit or blackmail people is explored.
Abstract
- Nowadays, people faced an emerging problem of AI synthesized face swapping videos, widely known as the Deep Fakes. These kinds of videos can be created to cause threats to privacy, fraudulence, and so on. Sometimes good quality DeepFake video recognition could be hard to distinguish with people's eyes. These can be used to manipulate public opinion during elections, commit fraud, and discredit or blackmail people. Therefore, there is a need for automated tools capable of detecting false multimedia content and avoiding the spread of dangerous false information. That's why researchers need to develop algorithms to detect them. One of the frequently studied deep learning techniques is GAN. These networks are commonly used to create DeepFake videos but are not used to detect them. Here exploring a solution based on a GAN discriminator to detect DeepFake videos. Train a GAN and extract the discriminator into a custom module for DeepFake detection. Test different discriminator architectures using different data sets to see how discriminator performance varies with different settings and training methods.