The identification of edited videos is a critical issue in multimedia forensics that has received a lot of attention in recent years. However, a frequent feature of published research is that forensic analysis is usually performed on data before it is used. This project tackles the difficult situation of altered videos that are uploaded on social media networks. In this regard, a large-scale performance evaluation using general-purpose deep learning has been carried out. The project shows that the output of differently trained networks can carry useful forensic information for the identification of the specific technique used for visual manipulation, both for shared and non-shared data.

• Our system will be able to generate a deepfake version of the user’s face in order to differentiate between their real and fake features
• Ethical wise, our database will be secure as in the consent of the user will be taken in consideration before running any process in our system.
• Our API will be generated based on taking the video/image as a resource and performing several action on it. The actions will be the preprocessing on videos/images and the detection.
• Our API will send a message to the user with the output of the classification that occurred, saying whether it’s real or fake.

• Our system is an API that helps in the detection of forged images and videos from surfacing the internet and spreading widely causing false information and rumors
• The system is made up of trained top CNN models on the latest datasets that include the newest deepfake methods