Few-Shot Adversarial Learning of Realistic Neural Talking Head Models

txt


Samsung’s research center at Moscow has developed a new AI that can create talking avatars of photos and paintings without using any 3D modeling.

Paper Abstract: Several recent works have shown how highly realistic human head images can be obtained by training convolutional neural networks to generate them. In order to create a personalized talking head model, these works require training on a large dataset of images of a single person. However, in many practical scenarios, such personalized talking head models need to be learned from a few image views of a person, potentially even a single image. Here, we present a system with such few-shot capability. It performs lengthy meta-learning on a large dataset of videos, and after that is able to frame few- and one-shot learning of neural talking head models of previously unseen people as adversarial training problems with high capacity generators and discriminators. Crucially, the system is able to initialize the parameters of both the generator and the discriminator in a person-specific way, so that training can be based on just a few images and done quickly, despite the need to tune tens of millions of parameters. We show that such an approach is able to learn highly realistic and personalized talking head models of new people and even portrait paintings.

Leave A Comment

Discover, Learn and Evaluate AI Companies and Solutions

Save content to your library

Save case studies, articles, blog posts and more. Curate your research library with content directly from AI companies.

Login with LinkedIn Login with Twitter

Sign in with Email

The latest updates from AI companies in your industry

Get a weekly newsletter with the latest posts directly from the AI companies. Follow companies to tailor your feed.

LATEST | POPULAR
IMAGE ANNOTATION

Get your image labeled efficiently and cost-effectively.

SPELL - HYPERPARAMETER SEARCHES

Optimize hyperparameters to improve the accuracy of your model with the Spell Hyper command

SPELL - JUPYTER NOTEBOOK SERVER - GPU POWERED

Collaborative Jupyter Notebook or JupyterLab workspace server with powerful GPUs

SPELL - DEEP LEARNING PLATFORM

The fastest and most powerful end-to-end platform for machine learning and deep learning.

PRICE OPTIMISATION

Drive top line and margin growth through AI price optimisation

Make sure your business and career keeps up with the changing world.
Sign up