Label-driven weakly-supervised learning for multimodal deformable image registration

TL;DR

This paper introduces a label-driven, weakly-supervised deep learning method for fast, accurate, and deformable multimodal medical image registration that bypasses traditional intensity-based measures.

Contribution

It proposes a novel weakly-supervised training framework using anatomical labels to learn 3D voxel correspondence for multimodal image registration.

Findings

Median target registration error of 4.2 mm on landmarks

Median Dice score of 0.88 on prostate glands

Achieves real-time, label-free registration during inference

Abstract

Spatially aligning medical images from different modalities remains a challenging task, especially for intraoperative applications that require fast and robust algorithms. We propose a weakly-supervised, label-driven formulation for learning 3D voxel correspondence from higher-level label correspondence, thereby bypassing classical intensity-based image similarity measures. During training, a convolutional neural network is optimised by outputting a dense displacement field (DDF) that warps a set of available anatomical labels from the moving image to match their corresponding counterparts in the fixed image. These label pairs, including solid organs, ducts, vessels, point landmarks and other ad hoc structures, are only required at training time and can be spatially aligned by minimising a cross-entropy function of the warped moving label and the fixed label. During inference, the trained network takes a new image pair to predict an optimal DDF, resulting in a fully-automatic, label-free, real-time and deformable registration. For interventional applications where large global transformation prevails, we also propose a neural network architecture to jointly optimise the global- and local displacements. Experiment results are presented based on cross-validating registrations of 111 pairs of T2-weighted magnetic resonance images and 3D transrectal ultrasound images from prostate cancer patients with a total of over 4000 anatomical labels, yielding a median target registration error of 4.2 mm on landmark centroids and a median Dice of 0.88 on prostate glands.