The objective of this project is to detect deep fake videos using a multi-view vision transformer (MViT). Deep fake technology uses AI and machine learning models to manipulate video and audio, creating realistic fake content. Detecting such videos is important in ensuring the integrity of visual content and combating misinformation.
We leverage the power of deep learning, using a pre-trained MViT model, fine-tuning it for the task of binary classification: real vs fake videos.