Reconstruction and Reenactment Separated Method for Realistic Gaussian Head
This work addresses the need for efficient and realistic avatar creation for applications like virtual reality or gaming, though it appears incremental as it builds on existing Gaussian head methods.
The paper tackles the problem of generating controllable 3D head avatars from a single portrait image by proposing a reconstruction and reenactment separated framework, achieving 90 FPS rendering at 512x512 resolution and outperforming state-of-the-art methods.
In this paper, we explore a reconstruction and reenactment separated framework for 3D Gaussians head, which requires only a single portrait image as input to generate controllable avatar. Specifically, we developed a large-scale one-shot gaussian head generator built upon WebSSL and employed a two-stage training approach that significantly enhances the capabilities of generalization and high-frequency texture reconstruction. During inference, an ultra-lightweight gaussian avatar driven by control signals enables high frame-rate rendering, achieving 90 FPS at a resolution of 512x512. We further demonstrate that the proposed framework follows the scaling law, whereby increasing the parameter scale of the reconstruction module leads to improved performance. Moreover, thanks to the separation design, driving efficiency remains unaffected. Finally, extensive quantitative and qualitative experiments validate that our approach outperforms current state-of-the-art methods.