a resource manual for deepfake detection by @imreallyjohn
A deepfake is content ( could be an image, video or even audio! ), generated by an artificial intelligence, that is authentic in the eyes of a human being. The word deepfake is a combination of the words “deep learning” and “fake,” and primarily relates to content generated by an artificial neural network, a branch of machine learning. The most common form of deepfakes involve the generation and manipulation of human imagery. [1]
Deepfakes are usually generated using some visual synthesis algorithms. For example, here is one sample which attempts to synthesize important features of humans like expressions, gaze, mouth, etc; along with some reenactment and replacement (also known as Face Swap) 
Some of the state-of-the-art datasets which synthesize deepfakes very accurately are,
- Celeb-DF a large-scale deepfake video dataset containing 5639 HQ Deepfake videos [CVPR 2020] Dataset Download Link
- FaceForensics++ another large-scale deepfake video dataset containing 3000+ deepfake videos Dataset Download Link
For more insights please refer to Literature Survey