An Unsupervised Learning Model for Deformable Medical Image Registration

TL;DR

This paper introduces a fast, unsupervised learning-based 3D medical image registration method using CNNs, achieving accuracy comparable to state-of-the-art techniques but with significantly improved speed, enabling more efficient medical image analysis.

Contribution

It presents a novel unsupervised CNN-based registration model that is faster and does not require ground truth data, advancing learning-based medical image registration.

Findings

Achieves registration accuracy comparable to state-of-the-art methods.

Operates orders of magnitude faster than traditional approaches.

Does not require supervised ground truth or landmarks.

Abstract

We present a fast learning-based algorithm for deformable, pairwise 3D medical image registration. Current registration methods optimize an objective function independently for each pair of images, which can be time-consuming for large data. We define registration as a parametric function, and optimize its parameters given a set of images from a collection of interest. Given a new pair of scans, we can quickly compute a registration field by directly evaluating the function using the learned parameters. We model this function using a convolutional neural network (CNN), and use a spatial transform layer to reconstruct one image from another while imposing smoothness constraints on the registration field. The proposed method does not require supervised information such as ground truth registration fields or anatomical landmarks. We demonstrate registration accuracy comparable to state-of-the-art 3D image registration, while operating orders of magnitude faster in practice. Our method promises to significantly speed up medical image analysis and processing pipelines, while facilitating novel directions in learning-based registration and its applications. Our code is available at https://github.com/balakg/voxelmorph .