Imagine looking into a mirror, but not to see your own face, instead looking in the eyes of Barack Obama or Angela Merkel. Your facial expressions are seamlessly transferred to that other person's face, all in real time. The TNG Hardware Hacking Team has managed to create a prototype that transfers faces in real time based on Deepfakes. Neural networks detect faces within a video input, then translate and integrate them back to the video output. Through this technique it is possible to project deceptively real imitations onto other people. For this purpose, we used a Keras trained autoencoder networks and various face recognition algorithms. In this talk, the speakers will give an entertaining and vivid introduction into the world of real-time Deepfakes. In doing so, they will particularly focus on the deep learning techniques used in this application.

This talk is co-presented along with Martin Fortsch