SurFhead: Affine Rig Blending for Geometrically Accurate 2D Gaussian Surfel Head Avatars
This addresses the need for geometrically accurate head avatars for applications like mesh reconstruction and relighting, representing a novel integration of classical graphics techniques with modern Gaussian primitives.
The paper tackles the problem of capturing intricate geometric details in head avatar rendering from RGB videos by proposing SurFhead, which uses 2D Gaussian surfels with affine transformations and deformation transfer to achieve state-of-the-art geometry reconstruction and high-fidelity rendering in extreme poses.
Recent advancements in head avatar rendering using Gaussian primitives have achieved significantly high-fidelity results. Although precise head geometry is crucial for applications like mesh reconstruction and relighting, current methods struggle to capture intricate geometric details and render unseen poses due to their reliance on similarity transformations, which cannot handle stretch and shear transforms essential for detailed deformations of geometry. To address this, we propose SurFhead, a novel method that reconstructs riggable head geometry from RGB videos using 2D Gaussian surfels, which offer well-defined geometric properties, such as precise depth from fixed ray intersections and normals derived from their surface orientation, making them advantageous over 3D counterparts. SurFhead ensures high-fidelity rendering of both normals and images, even in extreme poses, by leveraging classical mesh-based deformation transfer and affine transformation interpolation. SurFhead introduces precise geometric deformation and blends surfels through polar decomposition of transformations, including those affecting normals. Our key contribution lies in bridging classical graphics techniques, such as mesh-based deformation, with modern Gaussian primitives, achieving state-of-the-art geometry reconstruction and rendering quality. Unlike previous avatar rendering approaches, SurFhead enables efficient reconstruction driven by Gaussian primitives while preserving high-fidelity geometry.