Linear and Deformable Image Registration with 3D Convolutional Neural Networks

TL;DR

This paper introduces a novel 3D CNN architecture for medical image registration that combines linear and deformable registration, achieving near real-time performance and superior results on MRI lung registration tasks.

Contribution

The paper presents a modular deep learning framework that unifies linear and deformable registration with smooth displacement fields, outperforming existing elastic registration methods.

Findings

Superior performance on MRI lung registration

Effective in interstitial lung disease classification

Achieves near real-time processing

Abstract

Image registration and in particular deformable registration methods are pillars of medical imaging. Inspired by the recent advances in deep learning, we propose in this paper, a novel convolutional neural network architecture that couples linear and deformable registration within a unified architecture endowed with near real-time performance. Our framework is modular with respect to the global transformation component, as well as with respect to the similarity function while it guarantees smooth displacement fields. We evaluate the performance of our network on the challenging problem of MRI lung registration, and demonstrate superior performance with respect to state of the art elastic registration methods. The proposed deformation (between inspiration & expiration) was considered within a clinically relevant task of interstitial lung disease (ILD) classification and showed promising results.