Dynamic imaging using motion-compensated smoothness regularization on manifolds (MoCo-SToRM)

TL;DR

This paper presents an unsupervised deep learning method for motion-compensated dynamic MRI that models motion fields on a manifold, enabling high-quality lung imaging from limited data.

Contribution

It introduces a novel unsupervised deep manifold learning framework that jointly estimates motion fields and images in dynamic MRI using a motion manifold prior.

Findings

Effective motion compensation in lung MRI from few radial spokes

Improved image quality with joint motion and image estimation

Demonstrated on high-resolution free-breathing lung MRI

Abstract

We introduce an unsupervised deep manifold learning algorithm for motion-compensated dynamic MRI. We assume that the motion fields in a free-breathing lung MRI dataset live on a manifold. The motion field at each time instant is modeled as the output of a deep generative model, driven by low-dimensional time-varying latent vectors that capture the temporal variability. The images at each time instant are modeled as the deformed version of an image template using the above motion fields. The template, the parameters of the deep generator, and the latent vectors are learned from the k-t space data in an unsupervised fashion. The manifold motion model serves as a regularizer, making the joint estimation of the motion fields and images from few radial spokes/frame well-posed. The utility of the algorithm is demonstrated in the context of motion-compensated high-resolution lung MRI.