Abstract

Deep fakes refer to the fabricated contents, particularly images and videos, using machine learning and computer vision techniques. Even though there are vast possibilities for utilizing these technologies in exciting and positive ways, there also has been a great amount of concern about the misuse of them. Creating falsified images can be used for sabotages and exploitations against people or institutions in various shape such as in financial fraud, revenge pornography, fake news, propaganda, among others. Thus, developing counterfeit-proof systems has garnered a lot of attention in digital society. In this project, we briefly analyze the various deep learning models to solve deepfake detection problem. Specifically, convolutional neural networks and recurrent neural networks are combined and evaluated.

Results/Implementation/Project Description

Conclusion

[ PDF (optional) ] 

[ Slides Kickoff/Final (optional)]