MI image registration using prior knowledge

TL;DR

This paper extends the concept of Mutual Information for image registration by incorporating prior knowledge through Focused Mutual Information, improving registration accuracy in clinical applications like dental and implant monitoring.

Contribution

It introduces (Normalized) Focused Mutual Information, a novel approach that integrates prior knowledge into MI for enhanced image registration performance.

Findings

Successfully applied to dental restoration follow-up

Effective in cephalometry registration

Improves implant monitoring accuracy

Abstract

Subtraction of aligned images is a means to assess changes in a wide variety of clinical applications. In this paper we explore the information theoretical origin of Mutual Information (MI), which is based on Shannon's entropy.However, the interpretation of standard MI registration as a communication channel suggests that MI is too restrictive a criterion. In this paper the concept of Mutual Information (MI) is extended to (Normalized) Focussed Mutual Information (FMI) to incorporate prior knowledge to overcome some shortcomings of MI. We use this to develop new methodologies to successfully address specific registration problems, the follow-up of dental restorations, cephalometry, and the monitoring of implants.