Loading...

Deepfake Videos Detection through Deep Analysis of Artifacts of Images

Aghababaei Harandi, Ali | 2021

143 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54720 (05)
  4. University: Sharif University of Technology
  5. Department: Electrical Engineering
  6. Advisor(s): Ghaemmaghami, Shahrokh; Eghlidos, Taraneh
  7. Abstract:
  8. DeepFake is a type of forgery that uses deep learning algorithms to make changes to audio and video content that the audience is unable to detect. Nowadays, due to the threats posed by the use of DeepFake to move people's faces in video, researchers' attention has been drawn to designing methods to detect this type of forgery. Detection methods are usually classified into two types. The first case is the extraction of features to detect forgery distortions, for example, the extraction of facial orientations to detect inconsistencies. The second case is the use of deep learning networks for feature extraction and classification, of which the EfficientNet network is an example. Despite the good performance of these methods, weaknesses such as the selection of unstable features, low generalizability and a large number of training parameters remain. In this research, to improve the expressed weaknesses, we have proposed a combined method that includes pattern noise and blur type extraction by neural networks and statistical processing, training of separate classifiers for each feature and finally statistically combining their output. The simulation results show comparable performance in terms of AUC, improvement in terms of generalizability (23.9%) and reduction in the number of training parameters (70%) against the EfficientNet network. Also, to improve the generalizability of deep learning networks to different datasets, we have proposed a method based on adding trainable parameters to the data and train these parameters to transfer new data to target classes, which in terms of AUC (14%) versus method without training Face X-ray has improved
  9. Keywords:
  10. Deep Learning ; Generalization ; Blur Type ; Deep Fake ; Noise Pattern

 Digital Object List

 Bookmark

No TOC