STGA: Selective-Training Gaussian Head Avatars
This work addresses the challenge of creating realistic animated head avatars for applications like virtual reality or gaming, though it appears incremental as it builds on existing FLAME and Gaussian splatting frameworks.
The authors tackled the problem of enhancing detail realism in dynamic head Gaussian avatars by proposing a selective-training approach that optimizes different 3D Gaussian splats per frame, achieving better results with shorter training time compared to network-based methods and more realistic details than mesh-based methods.
We propose selective-training Gaussian head avatars (STGA) to enhance the details of dynamic head Gaussian. The dynamic head Gaussian model is trained based on the FLAME parameterized model. Each Gaussian splat is embedded within the FLAME mesh to achieve mesh-based animation of the Gaussian model. Before training, our selection strategy calculates the 3D Gaussian splat to be optimized in each frame. The parameters of these 3D Gaussian splats are optimized in the training of each frame, while those of the other splats are frozen. This means that the splats participating in the optimization process differ in each frame, to improve the realism of fine details. Compared with network-based methods, our method achieves better results with shorter training time. Compared with mesh-based methods, our method produces more realistic details within the same training time. Additionally, the ablation experiment confirms that our method effectively enhances the quality of details.