PILOT: Physics-Informed Learned Optimized Trajectories for Accelerated MRI

Abstract

Magnetic Resonance Imaging (MRI) has long been considered to be among "the gold standards" of diagnostic medical imaging. The long acquisition times, however, render MRI prone to motion artifacts, let alone their adverse contribution to the relative high costs of MRI examination. Over the last few decades, multiple studies have focused on the development of both physical and post-processing methods for accelerated acquisition of MRI scans. These two approaches, however, have so far been addressed separately. On the other hand, recent works in optical computational imaging have demonstrated growing success of concurrent learning-based design of data acquisition and image reconstruction schemes. Such schemes have already demonstrated substantial effectiveness, leading to considerably shorter acquisition times and improved quality of image reconstruction. Inspired by this initial success, in this work, we propose a novel approach to the learning of optimal schemes for conjoint acquisition and reconstruction of MRI scans, with the optimization carried out simultaneously with respect to the time-efficiency of data acquisition and the quality of resulting reconstructions. To be of a practical value, the schemes are encoded in the form of general k-space trajectories, whose associated magnetic gradients are constrained to obey a set of predefined hardware requirements (as defined in terms of, e.g., peak currents and maximum slew rates of magnetic gradients). With this proviso in mind, we propose a novel algorithm for the end-to-end training of a combined acquisition-reconstruction pipeline using a deep neural network with differentiable forward-and backpropagation operators. We also demonstrate the effectiveness of the proposed solution in application to both image reconstruction and image segmentation, reporting substantial improvements in terms of acceleration factors as well as the quality of these end tasks.Code available at https://github.com/tomer196/PILOT.

J-MoDL: Joint Model-Based Deep Learning for Optimized Sampling and Reconstruction

“…The fast image recovery offered by deep learning methods such as LOUPE [20] and PILOT [22] offer an alternative to accelerate (8) and simultaneously solve for the hyperparameters Φ. Instead of directly solving for the k-space locations, the LOUPE approach optimizes for the sampling density [20].…”

“…Note that this is a relaxation of the original problem. Since the density may not capture the dependencies between k-space samples (e.g., conjugate symmetry when the image is real or smoothly varying phase), improved gains may be obtained by directly solving for the k-space sampling locations rather than the density, such that the 2 training error is minimized in PILOT [22] {Θ * , Φ * } = arg min…”

“…conjugate symmetry) between the k-space samples. The approaches in [21], [22] instead consider a sampling mask to choose a subset of Cartesian samples, which is optimized. A challenge with [22] is that the final sampling pattern heavily depends on the initialization of the pattern.…”

“…The approaches in [21], [22] instead consider a sampling mask to choose a subset of Cartesian samples, which is optimized. A challenge with [22] is that the final sampling pattern heavily depends on the initialization of the pattern. This may be attributed to the complex nature of the optimization landscape, consisting of several local minima.…”

“…The main focus of this work is to jointly optimize the sampling pattern and the deep network parameters for parallel MRI reconstruction. Most of the previous optimization strategies [10], [11], [20]- [22] are only restricted to the single-channel setting. We rely on an algorithm-dependent strategy to search for the best sampling pattern in the multichannel setting.…”