SShaDe: scalable shape deformation via local representations

TL;DR

This work proposes a novel approach for tackling mesh deformation tasks on high-resolution meshes by reducing the input size with a fast remeshing technique and preserving a consistent representation of the original mesh with local reference frames, which provides a solution that is both scalable and robust in multiple applications.

Abstract

With the increase in computational power for the available hardware, the demand for high-resolution data in computer graphics applications increases. Consequently, classical geometry processing techniques based on linear algebra solutions are starting to become obsolete. In this setting, we propose a novel approach for tackling mesh deformation tasks on high-resolution meshes. By reducing the input size with a fast remeshing technique and preserving a consistent representation of the original mesh with local reference frames, we provide a solution that is both scalable and robust in multiple applications, such as as-rigid-as-possible deformations, non-rigid isometric transformations, and pose transfer tasks. We extensively test our technique and compare it against state-of-the-art methods, proving that our approach can handle meshes with hundreds of thousands of vertices in tens of seconds while still achieving results comparable with the other solutions.

Figures (9)

• Figure 1: The vertices of two Voronoi regions on the high-resolution shape (solid) are projected onto the low-resolution representation (dashed, shaded in yellow) with the approach from maggioli:2023:rematching (left) and our method (right). The color of the vertices (dark red or dark blue) reveals in which part of the low-resolution surface they are projected (respectively, light red or light blue).
• Figure 2: An example of local reference frame at a triangle (left) and how it is affected by a rigid transformation (center) and a non-rigid deformation (right) of the surface. The blue vector is always aligned with the surface normal, the red vector points towards the snout, and the green vector towards the left of the dog's head.
• Figure 3: Comparison between the direct application of our approach (left) and the averaging of multiple iterations (right). The closeups on the arms show the difference in smoothness between the two approaches.
• Figure 4: Comparison of our solution with the baseline approaches on sample deformations from the DeFAUST dataset. The close-ups show the deformation artifacts, and the edge error as in \ref{['eq:edge-error']} is shown as a hot colormap.
• Figure 5: Comparison of our solution with the baseline approaches on some user-defined deformations from baieri:2024:implicitarap.