Volumetric Avatar

Volumetric avatars are 3D digital representations of humans, aiming for photorealistic rendering and dynamic animation for applications like AR/VR and video conferencing. Current research focuses on improving the accuracy and efficiency of these avatars, particularly in challenging areas like facial features and clothing, using neural radiance fields (NeRFs) and point-based representations often guided by skeletal information or UV maps. These advancements enable more realistic and controllable avatars, with a particular emphasis on local geometry and texture editing, improving the quality of novel view synthesis and pose rendering for various applications.