Robust Biharmonic Skinning Using Geometric Fields

A new skinning method replaces the usual tetrahedral volume step with a mesh-free formulation built on geometric fields. That matters because bounded biharmonic weights are widely used for character rigging, reduced-order simulation, and geometry descriptors, but they often break when the surface can’t be cleanly tetrahedralized.

The key practical win is robustness: the method is designed to work on open surfaces, triangle soups, and point clouds, where conventional bounded biharmonic approaches can fail outright. It uses hardware ray tracing to support a specialized Lagrangian representation, letting the solver optimize biharmonic energy and boundary conditions without relying on finite elements.

For game teams, this is interesting anywhere automatic rigging or deformation tools need to survive messy real-world assets. The formulation also keeps room for artist control through weight painting during optimization, which is important if the goal is to slot into production rather than replace it with a black box.

The work comes from Ana Dodik, Vincent Sitzmann, Justin Solomon, and Oded Stein, and it has been revised multiple times since its first submission on June 1, 2024, with a later version posted on August 3, 2026. A thorough qualitative and quantitative evaluation accompanies the method, suggesting the authors are positioning it as a practical alternative rather than a narrow research demo.