Deepfakes in Realtime - How Neural Networks Are Changing Our World

Go beyond simple face swaps. Learn to build a real-time deepfake pipeline using GANs for photorealistic results.

#1about 4 minutes

The danger and deception of deepfake technology

An introductory example using a fake Obama video highlights the potential for misinformation and the importance of public awareness.

#2about 4 minutes

Differentiating fakes from true deepfakes

A breakdown of related but distinct technologies like lip syncing, face swapping, and real-time reenactment clarifies what constitutes a genuine deepfake.

#3about 13 minutes

How traditional deepfakes are made with Deep Face Lab

The original deepfake workflow involves preparing a dataset with FaceNet, training a shared autoencoder, and then running inference to generate the final video.

#4about 10 minutes

Building a real-time deepfake pipeline

To achieve real-time performance, the original pipeline is modified with faster face segmentation using MobileNet and U-Net, and image inpainting to remove the original head.

#5about 8 minutes

Using GANs to improve deepfake image quality

Generative Adversarial Networks (GANs) are introduced to the training process, where a generator and a discriminator compete to produce more realistic and artifact-free faces.

#6about 2 minutes

Live demo and limitations of data-driven models

A live demonstration shows the real-time head replacement and reveals a key limitation where the model can only reproduce expressions seen in its training data.

#7about 4 minutes

Applications in entertainment and detecting fake news

Deepfakes offer a cost-effective alternative to CGI in the movie industry and can also be used to train discriminator models for detecting fake news.

#8about 2 minutes

The future of deepfakes beyond face swapping

The next evolution of deepfake technology will likely involve higher resolutions, full-body manipulation, and real-time voice cloning to create more comprehensive fakes.