HAHA: Highly Articulated Gaussian Human Avatars with Textured Mesh Prior
This addresses efficient and high-fidelity avatar animation for applications like virtual reality or gaming, though it is incremental as it builds on existing Gaussian splatting and mesh-based techniques.
The paper tackles animatable human avatar generation from monocular videos by balancing Gaussian splatting and textured mesh rendering, achieving state-of-the-art quality on SnapshotPeople with less than a third of Gaussians and outperforming prior methods on novel poses in X-Humans.
We present HAHA - a novel approach for animatable human avatar generation from monocular input videos. The proposed method relies on learning the trade-off between the use of Gaussian splatting and a textured mesh for efficient and high fidelity rendering. We demonstrate its efficiency to animate and render full-body human avatars controlled via the SMPL-X parametric model. Our model learns to apply Gaussian splatting only in areas of the SMPL-X mesh where it is necessary, like hair and out-of-mesh clothing. This results in a minimal number of Gaussians being used to represent the full avatar, and reduced rendering artifacts. This allows us to handle the animation of small body parts such as fingers that are traditionally disregarded. We demonstrate the effectiveness of our approach on two open datasets: SnapshotPeople and X-Humans. Our method demonstrates on par reconstruction quality to the state-of-the-art on SnapshotPeople, while using less than a third of Gaussians. HAHA outperforms previous state-of-the-art on novel poses from X-Humans both quantitatively and qualitatively.