Neural head avatars offer a new fascinating way of creating virtual head models. They bypass the complexity of realistic physics-based modeling of human avatars by learning the shape and appearance directly from the videos of talking people. In this recent report, the authors present the first system capable of creating megapixel avatars from single portrait images. According to the report, the author's method outperforms its competitors in the quality of the cross-driving results and manages to preserve the high-resolution appearance of the source image even for out-of-domain examples like paintings.

The post Captivatingly Terrifying? One-Shot Megapixel Neural Head Avatars appeared first on ComplexDiscovery.