Neural Subdivision
This addresses the problem of flexible and generalizable geometry modeling for computer graphics and 3D modeling applications, though it is incremental as it builds on classical subdivision techniques with neural enhancements.
The paper tackles the problem of data-driven coarse-to-fine geometry modeling by introducing Neural Subdivision, a framework that recursively subdivides coarse triangle meshes using a neural network to predict vertex positions, enabling learning of complex non-linear schemes beyond classical linear averaging. The result is a method that generalizes across discretizations, preserves manifold structure, and can generate reasonable subdivisions for novel shapes even when trained on a single high-resolution mesh.
This paper introduces Neural Subdivision, a novel framework for data-driven coarse-to-fine geometry modeling. During inference, our method takes a coarse triangle mesh as input and recursively subdivides it to a finer geometry by applying the fixed topological updates of Loop Subdivision, but predicting vertex positions using a neural network conditioned on the local geometry of a patch. This approach enables us to learn complex non-linear subdivision schemes, beyond simple linear averaging used in classical techniques. One of our key contributions is a novel self-supervised training setup that only requires a set of high-resolution meshes for learning network weights. For any training shape, we stochastically generate diverse low-resolution discretizations of coarse counterparts, while maintaining a bijective mapping that prescribes the exact target position of every new vertex during the subdivision process. This leads to a very efficient and accurate loss function for conditional mesh generation, and enables us to train a method that generalizes across discretizations and favors preserving the manifold structure of the output. During training we optimize for the same set of network weights across all local mesh patches, thus providing an architecture that is not constrained to a specific input mesh, fixed genus, or category. Our network encodes patch geometry in a local frame in a rotation- and translation-invariant manner. Jointly, these design choices enable our method to generalize well, and we demonstrate that even when trained on a single high-resolution mesh our method generates reasonable subdivisions for novel shapes.