Metric Registration of Curves and Surfaces using Optimal Control

TL;DR

This paper unifies shape analysis of curves and surfaces through Riemannian metrics, simplifying geodesic computations by formulating shape matching as optimal control problems with relaxed boundary conditions.

Contribution

It introduces a unified framework for shape metrics via Riemannian submersions and develops an inexact, optimal control-based approach for shape matching problems.

Findings

Unified shape metrics through Riemannian submersions.

Simplified geodesic computations using chordal distances.

Framework applicable to various shape metric models.

Abstract

This paper presents an overview of recent developments in the analysis of shapes such as curves and surfaces through Riemannian metrics. We show that several constructions of metrics on spaces of submanifolds can be unified through the prism of Riemannian submersions, with shape space metrics being induced from metrics defined on the top spaces. Computing the resulting Riemannian distances involves solving geodesic matching problems with boundary conditions. To deal efficiently with such variational problems, one can rely on an auxiliary family of "chordal" distances to simplify the treatment of boundary conditions, which we use to come up with a relaxed inexact formulation of the matching problem. This also allows to turn shape matching into optimal control problems and give a common framework to address them in practice. We then specify our analysis to the cases of intrinsic shape metrics defined using invariant Sobolev metrics on parametrized immersions, outer shape metrics induced from metrics on diffeomorphism groups of the ambient space and finally a recent hybrid model that combines those two approaches.