Multiresolution Search of the Rigid Motion Space for Intensity Based Registration

TL;DR

This paper investigates how low-resolution intensity-based registration can effectively bound high-resolution registration, enabling more efficient multiresolution algorithms for 2D and 3D image registration and symmetry detection.

Contribution

It establishes a theoretical relationship between low and high-resolution target functions and integrates this into a multiresolution registration framework with demonstrated efficiency gains.

Findings

Low-resolution target values tightly bound high-resolution targets in natural images.

Multiresolution approach improves efficiency of global optimization in registration tasks.

Method successfully applied to 2D/3D registration and symmetry detection.

Abstract

We study the relation between the target functions of low-resolution and high-resolution intensity-based registration for the class of rigid transformations. Our results show that low resolution target values can tightly bound the high-resolution target function in natural images. This can help with analyzing and better understanding the process of multiresolution image registration. It also gives a guideline for designing multiresolution algorithms in which the search space in higher resolution registration is restricted given the fitness values for lower resolution image pairs. To demonstrate this, we incorporate our multiresolution technique into a Lipschitz global optimization framework. We show that using the multiresolution scheme can result in large gains in the efficiency of such algorithms. The method is evaluated by applying to 2D and 3D registration problems as well as the detection of reflective symmetry in 2D and 3D images.