GraphAvatar: Compact Head Avatars with GNN-Generated 3D Gaussians

TL;DR

GraphAvatar introduces a GNN-based approach to generate compact, high-fidelity head avatars using 3D Gaussians, significantly reducing storage needs while improving rendering quality.

Contribution

It proposes a novel GNN-driven method to generate 3D Gaussians for avatars, drastically reducing storage overhead and enhancing rendering fidelity compared to prior methods.

Findings

Surpasses existing methods in visual fidelity

Reduces storage to just 10MB

Effective in refining face-tracking errors

Abstract

Rendering photorealistic head avatars from arbitrary viewpoints is crucial for various applications like virtual reality. Although previous methods based on Neural Radiance Fields (NeRF) can achieve impressive results, they lack fidelity and efficiency. Recent methods using 3D Gaussian Splatting (3DGS) have improved rendering quality and real-time performance but still require significant storage overhead. In this paper, we introduce a method called GraphAvatar that utilizes Graph Neural Networks (GNN) to generate 3D Gaussians for the head avatar. Specifically, GraphAvatar trains a geometric GNN and an appearance GNN to generate the attributes of the 3D Gaussians from the tracked mesh. Therefore, our method can store the GNN models instead of the 3D Gaussians, significantly reducing the storage overhead to just 10MB. To reduce the impact of face-tracking errors, we also present a novel graph-guided optimization module to refine face-tracking parameters during training. Finally, we introduce a 3D-aware enhancer for post-processing to enhance the rendering quality. We conduct comprehensive experiments to demonstrate the advantages of GraphAvatar, surpassing existing methods in visual fidelity and storage consumption. The ablation study sheds light on the trade-offs between rendering quality and model size. The code will be released at: https://github.com/ucwxb/GraphAvatar