Facial Deepfake Detection Based on Spactial Image Features

With the continuous development of deepfake technology, led by "ai in face" many depth of forged related technology began to fill in the line of sight of the public, its not only in the entertainment industry provides extremely convenient function, but also brought including but not limited to information fraud many malignant events, so the depth of forged detection technology is particularly important. This article begins by reviewing the history of deep forgery, And the most common reproduction, editing, replacement and synthesis, Following up with examples and comparisons of data sets emerging in recent years, Then, the face depth forgery detection technology based on spatial image features based on attention network, based on autoencoder, based on vision Transformer and based on data enhancement of four directions are roughly introduced, Finally, the paper explores some of the problems in the field, and with the existing challenges, Some future development directions are proposed, The full text is also summarized.