English

Deep neural networks on diffeomorphism groups for optimal shape reparameterization

Optimization and Control 2023-11-14 v2 Machine Learning Differential Geometry

Abstract

One of the fundamental problems in shape analysis is to align curves or surfaces before computing geodesic distances between their shapes. Finding the optimal reparametrization realizing this alignment is a computationally demanding task, typically done by solving an optimization problem on the diffeomorphism group. In this paper, we propose an algorithm for constructing approximations of orientation-preserving diffeomorphisms by composition of elementary diffeomorphisms. The algorithm is implemented using PyTorch, and is applicable for both unparametrized curves and surfaces. Moreover, we show universal approximation properties for the constructed architectures, and obtain bounds for the Lipschitz constants of the resulting diffeomorphisms.

Keywords

Cite

@article{arxiv.2207.11141,
  title  = {Deep neural networks on diffeomorphism groups for optimal shape reparameterization},
  author = {Elena Celledoni and Helge Glöckner and Jørgen Riseth and Alexander Schmeding},
  journal= {arXiv preprint arXiv:2207.11141},
  year   = {2023}
}

Comments

36 pages, 11 figures. Accepted by BIT Numerical Mathematics, not yet published

R2 v1 2026-06-25T01:09:01.521Z