Machine learning researchers have produced a system that can recreate lifelike motion from just a single frame of a person’s face, opening up the possibility of animating not just photos but also paintings. It’s not perfect, but when it works, it is — like much AI work these days — eerie and fascinating.
The model is documented in a paper published by Samsung AI Center, which you can read here on Arxiv. It’s a new method of applying facial landmarks on a source face — any talking head will do — to the facial data of a target face, making the target face do what the source face does.
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.