This technology is called "MegaPortraits". The avatars are designed to render the head in greater detail and are capable of capturing facial expressions and other subtle details, creating realistic digital avatars that can be used in digital media, virtual reality, augmented reality and other applications. The system is capable of generating high-quality avatars from a single photograph, making it suitable for applications where a large number of avatars are needed and the time to generate them is limited. The system can also learn from multiple photographs, allowing it to generate more natural-looking avatars.