Deep Generative Adversarial Neural Networks for Compressive Sensing MRI

Abstract

Undersampled magnetic resonance image (MRI) reconstruction is typically an ill-posed linear inverse task. The time and resource intensive computations require trade offs between accuracy and speed. In addition, state-of-the-art compressed sensing (CS) analytics are not cognizant of the image diagnostic quality. To address these challenges, we propose a novel CS framework that uses generative adversarial networks (GAN) to model the (low-dimensional) manifold of high-quality MR images. Leveraging a mixture of least-squares (LS) GANs and pixel-wise ℓ1/ℓ2 cost, a deep residual network with skip connections is trained as the generator that learns to remove the aliasing artifacts by projecting onto the image manifold. The LSGAN learns the texture details, while the ℓ1/ℓ2 cost suppresses high-frequency noise. A discriminator network, which is a multilayer convolutional neural network (CNN), plays the role of a perceptual cost that is then jointly trained based on high quality MR images to score the quality of retrieved images. In the operational phase, an initial aliased estimate (e.g., simply obtained by zero-filling) is propagated into the trained generator to output the desired reconstruction. This demands very low computational overhead. Extensive evaluations are performed on a large contrast-enhanced MR dataset of pediatric patients. Images rated by expert radiologists corroborate that GANCS retrieves higher quality images with improved fine texture details compared with conventional Wavelet-based and dictionary-learning based CS schemes as well as with deeplearning based schemes using pixel-wise training. In addition, it offers reconstruction times of under a few milliseconds, which is two orders of magnitude faster than current state-of-the-art CS-MRI schemes.

Real‐time cardiovascular MR with spatio‐temporal artifact suppression using deep learning–proof of concept in congenital heart disease

“…Finally, in this study, we choose to minimize the ‐loss and do not fully explore other loss functions such as ‐loss or SSIM‐loss. The ‐loss function has been shown to provide superior training or improved results in image restoration, super resolution, and MR image reconstruction . Therefore, we performed preliminary tests comparing the ‐loss and ‐loss using the tGA rot trajectory.…”

Deep residual network for off‐resonance artifact correction with application to pediatric body MRA with 3D cones

“…Training Off‐ResNet with a generative adversarial network could be a way to address this issue. It has been noted that generative adversarial networks can improve perceptual image quality and are better at maintaining high‐frequency structures than using only a data consistency loss . We have proposed a theoretically motivated architecture.…”

“…There have been many published techniques on undersampled image reconstruction, and image reconstruction can also be interpreted as an image artifact correction problem, suggesting feasibility for deblurring off‐resonance. Some convolutional neural network techniques operate entirely in the image domain and enhance image quality with supervised, perceptual, and adversarial losses . Still operating primarily in the image domain but drawing from a SENSE‐based reconstruction, there are techniques that use deep variational networks to learn a deep model prior to replacing the sparsity regularizers in the compressed‐sensing reconstruction equation .…”

Highly undersampled magnetic resonance imaging reconstruction using autoencoding priors

“…The first category is supervised learning approaches, which usually need an input‐output pair in the training stage and start from the same type of input data in the testing stage. There are roughly two types of supervised learning methods: One is data‐driven, which collects a large size of a training set to train a network that maps the observed data and ideal reconstruction purely from the perspective of end‐to‐end learning . The representative and widely used method of this kind is the convolution neural network (CNN) MRI .…”

“…The main contributions of this work are as follows: To our best knowledge, this is the first work to introduce the DAE prior for MRI reconstruction. Unlike the recent deep CNN‐based methods using an end‐to‐end learning fashion, we use network learning as a tool to learn general prior information and incorporate it into the constrained reconstruction framework, whose flowchart illustration is shown in Figure . Once the network‐learned image prior is obtained, it can be applied to the reconstruction tasks with different sampling trajectories and acceleration factors, and can guarantee promising results. More important, two advanced strategies are proposed to enhance the naïve DAE prior, termed EDAEP (enhanced DAEP).…”