Framing image registration as a landmark detection problem for label-noise-aware task representation (HitR)

TL;DR

This paper introduces Landmark Hit Rate (HitR), a new evaluation metric for image registration that accounts for clinical relevance and annotation noise, providing more meaningful assessments for biomedical applications.

Contribution

The study proposes HitR, a label-noise-aware evaluation metric based on confidence zones derived from inter-rater variance, shifting the focus to clinical relevance in registration assessment.

Findings

HitR effectively captures registration success within clinically meaningful zones.

It accounts for annotation noise, making evaluations more realistic.

Performance curves enable task-specific accuracy assessment.

Abstract

Accurate image registration is pivotal in biomedical image analysis, where selecting suitable registration algorithms demands careful consideration. While numerous algorithms are available, the evaluation metrics to assess their performance have remained relatively static. This study addresses this challenge by introducing a novel evaluation metric termed Landmark Hit Rate (HitR), which focuses on the clinical relevance of image registration accuracy. Unlike traditional metrics such as Target Registration Error, which emphasize subresolution differences, HitR considers whether registration algorithms successfully position landmarks within defined confidence zones. This paradigm shift acknowledges the inherent annotation noise in medical images, allowing for more meaningful assessments. To equip HitR with label-noise-awareness, we propose defining these confidence zones based on an Inter-rater Variance analysis. Consequently, hit rate curves are computed for varying landmark zone sizes, enabling performance measurement for a task-specific level of accuracy. Our approach offers a more realistic and meaningful assessment of image registration algorithms, reflecting their suitability for clinical and biomedical applications.