Instant Volumetric Head Avatars
This enables rapid creation of interactive avatars for applications like virtual reality or gaming, though it is incremental as it builds on existing neural radiance field and parametric face model techniques.
The paper tackles the problem of slow training times for photo-realistic digital head avatars by introducing INSTA, which reconstructs avatars in less than 10 minutes on modern GPUs, orders of magnitude faster than state-of-the-art methods, while maintaining high rendering quality.
We present Instant Volumetric Head Avatars (INSTA), a novel approach for reconstructing photo-realistic digital avatars instantaneously. INSTA models a dynamic neural radiance field based on neural graphics primitives embedded around a parametric face model. Our pipeline is trained on a single monocular RGB portrait video that observes the subject under different expressions and views. While state-of-the-art methods take up to several days to train an avatar, our method can reconstruct a digital avatar in less than 10 minutes on modern GPU hardware, which is orders of magnitude faster than previous solutions. In addition, it allows for the interactive rendering of novel poses and expressions. By leveraging the geometry prior of the underlying parametric face model, we demonstrate that INSTA extrapolates to unseen poses. In quantitative and qualitative studies on various subjects, INSTA outperforms state-of-the-art methods regarding rendering quality and training time.