3D Gaussian Blendshapes for Head Avatar Animation

TL;DR

This paper presents a novel 3D Gaussian blendshape method for creating photorealistic head avatars from monocular videos, enabling real-time high-fidelity animation with improved detail capture and rendering efficiency.

Contribution

The introduction of 3D Gaussian blendshapes for head avatars offers a new way to model expressions and appearance with high detail and real-time performance, surpassing previous methods.

Findings

Better high-frequency detail capture than state-of-the-art methods.

Achieves real-time high-fidelity head avatar animation.

Provides superior rendering performance.

Abstract

We introduce 3D Gaussian blendshapes for modeling photorealistic head avatars. Taking a monocular video as input, we learn a base head model of neutral expression, along with a group of expression blendshapes, each of which corresponds to a basis expression in classical parametric face models. Both the neutral model and expression blendshapes are represented as 3D Gaussians, which contain a few properties to depict the avatar appearance. The avatar model of an arbitrary expression can be effectively generated by combining the neutral model and expression blendshapes through linear blending of Gaussians with the expression coefficients. High-fidelity head avatar animations can be synthesized in real time using Gaussian splatting. Compared to state-of-the-art methods, our Gaussian blendshape representation better captures high-frequency details exhibited in input video, and achieves superior rendering performance.