Deepfake Detection in the Age of Synthetic Media: A Systematic Review

In the current generation social media is the main form of communication for people around the globe. But it is useful in a positive manner. Also, it has some negative impacts such as the spreading of fake news by editing images and videos and sharing them on social media. This is the main area the entire world is looking at because of Artificial intelligence it has become much easier than before to create fake images and videos by using this. This creation of fake images or videos is termed Deepfake. These are the major threats to celebrities and politicians by changing the images and the audio of their videos and by spreading fake news and making people panic. To detect deepfakes many machine learning algorithms use convolutional neural networks which is used for detecting fake images by image preprocessing and recurrent neural networks are helpful in audio processing. This paper provides an exclusive study on deepfake detection mechanisms and also helpful in identifying the model with better accuracy helpful in making an hybrid model by the combination of the models to achieve the accuracy and better results.