Research on 3D Driveable Digital Human Generation System

Abstract

The Metaverse has been gaining more attention from researchers worldwide. As the avatar of human in the Metaverse, three dimensional (3D) digital human has also attracted extensive attention. At present, the large-scale application of 3D digital human cannot be realized. The reason is that the construction cost of 3D avatar is too high, and it is mostly completed manually, especially for building a drivable 3D digital human. At present, the research on intelligent modeling of 3D digital human is modularized, lacking system integration to generate a drivable 3D digital human. This paper proposes a 3D drivable digital human generation method, and constructs an end-to-end testing system. The test proves that it can realize the generation of 3D drivable digital human based on single-view RGB image, and realize the driven of it by video from ordinary camera.