Two-stage deep learning for accelerated 3D time-of-flight MRA without matched training data

Chung, Hyungjin; Cha, Eunju; Sunwoo, Leonard; Ye, Jong Chul

doi:10.1016/j.media.2021.102047

Cited by 19 publications

(13 citation statements)

References 49 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Therefore, we need to train the reconstruction network such that the network learns the inverse mapping with respect to the forward operation T n F L ψ . For training, we utilize physics-informed OT-cycleGAN, which was shown to be especially effective for MRI reconstruction (Oh et al, 2020b;Chung et al, 2021;. Specifically, let X and Z denote the domains for the motion artifact free high resolution images and the aliased images from the forward operation in (8), respectively.…”

Section: Bootstrap Aggregation For Artifact Correction and Super Reso...mentioning

confidence: 99%

“…where Π(µ, ν) denotes the set of the joint distribution whose marginals are µ and ν. The geometric meaning of (10) was discussed in detail in (Oh et al, 2020b;Chung et al, 2021;, which aims to minimize the statistical distances between transported and the empirical measures in X and Z simultaneously. Furthermore, the primal formulation of the unsupervised learning in (10) can be represented by a dual formulation:…”

Section: Bootstrap Aggregation For Artifact Correction and Super Reso...mentioning

confidence: 99%

See 1 more Smart Citation

Simultaneous super-resolution and motion artifact removal in diffusion-weighted MRI using unsupervised deep learning

Chung¹,

Kim²,

Yoon³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Diffusion-weighted MRI is nowadays performed routinely due to its prognostic ability, yet the quality of the scans are often unsatisfactory which can subsequently hamper the clinical utility. To overcome the limitations, here we propose a fully unsupervised quality enhancement scheme, which boosts the resolution and removes the motion artifact simultaneously. This process is done by first training the network using optimal transport driven cycleGAN with stochastic degradation block which learns to remove aliasing artifacts and enhance the resolution, then using the trained network in the test stage by utilizing bootstrap subsampling and aggregation for motion artifact suppression. We further show that we can control the trade-off between the amount of artifact correction and resolution by controlling the bootstrap subsampling ratio at the inference stage. To the best of our knowledge, the proposed method is the first to tackle super-resolution and motion artifact correction simultaneously in the context of MRI using unsupervised learning. We demonstrate the efficiency of our method by applying it to both quantitative evaluation using simulation study, and to in vivo diffusion-weighted MR scans, which shows that our method is superior to the current state-of-the-art methods. The proposed method is flexible in that it can be applied to various quality enhancement schemes in other types of MR scans, and also directly to the quality enhancement of apparent diffusion coefficient maps.

show abstract

Section: Bootstrap Aggregation For Artifact Correction and Super Reso...mentioning

confidence: 99%

Section: Bootstrap Aggregation For Artifact Correction and Super Reso...mentioning

confidence: 99%

Simultaneous super-resolution and motion artifact removal in diffusion-weighted MRI using unsupervised deep learning

Chung¹,

Kim²,

Yoon³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…In many inverse problems, additional regularization is often used. For example, one could use the following [23], [24]:…”

Section: A Derivation Of Cycle-free Cycleganmentioning

confidence: 99%

“…Note that the first two terms in ( 15) are computed by using both x and y, whereas the last term is only with respect to y. From the optimal transport perspective, this makes a huge differences, since the computation of the last term is trivial whereas the first term requires the dual formulation [23], [24]. One of the most important contributions of our companion paper [10] is to show that the primal formulation of the unsupervised learning in (13) with the transport cost (15) can be represented by:…”

Section: A Derivation Of Cycle-free Cycleganmentioning

confidence: 99%

See 1 more Smart Citation

Cycle-free CycleGAN using Invertible Generator for Unsupervised Low-Dose CT Denoising

Kwon¹,

Ye²

2021

Preprint

Self Cite

View full text Add to dashboard Cite

Recently, CycleGAN was shown to provide highperformance, ultra-fast denoising for low-dose X-ray computed tomography (CT) without the need for a paired training dataset. Although this was possible thanks to cycle consistency, CycleGAN requires two generators and two discriminators to enforce cycle consistency, demanding significant GPU resources and technical skills for training. A recent proposal of tunable CycleGAN with Adaptive Instance Normalization (AdaIN) alleviates the problem in part by using a single generator. However, two discriminators and an additional AdaIN code generator are still required for training. To solve this problem, here we present a novel cyclefree Cycle-GAN architecture, which consists of a single generator and a discriminator but still guarantees cycle consistency. The main innovation comes from the observation that the use of an invertible generator automatically fulfills the cycle consistency condition and eliminates the additional discriminator in the CycleGAN formulation. To make the invertible generator more effective, our network is implemented in the wavelet residual domain. Extensive experiments using various levels of low-dose CT images confirm that our method can significantly improve denoising performance using only 10% of learnable parameters and faster training time compared to the conventional CycleGAN.

show abstract

Knowledge‐driven deep learning for fast MR imaging: Undersampled MR image reconstruction from supervised to un‐supervised learning

Wang,

Wu,

Jia

et al. 2024

Magnetic Resonance in Med

View full text Add to dashboard Cite

Deep learning (DL) has emerged as a leading approach in accelerating MRI. It employs deep neural networks to extract knowledge from available datasets and then applies the trained networks to reconstruct accurate images from limited measurements. Unlike natural image restoration problems, MRI involves physics‐based imaging processes, unique data properties, and diverse imaging tasks. This domain knowledge needs to be integrated with data‐driven approaches. Our review will introduce the significant challenges faced by such knowledge‐driven DL approaches in the context of fast MRI along with several notable solutions, which include learning neural networks and addressing different imaging application scenarios. The traits and trends of these techniques have also been given which have shifted from supervised learning to semi‐supervised learning, and finally, to unsupervised learning methods. In addition, MR vendors' choices of DL reconstruction have been provided along with some discussions on open questions and future directions, which are critical for the reliable imaging systems.

show abstract

Two-stage deep learning for accelerated 3D time-of-flight MRA without matched training data

Cited by 19 publications

References 49 publications

Simultaneous super-resolution and motion artifact removal in diffusion-weighted MRI using unsupervised deep learning

Simultaneous super-resolution and motion artifact removal in diffusion-weighted MRI using unsupervised deep learning

Cycle-free CycleGAN using Invertible Generator for Unsupervised Low-Dose CT Denoising

Knowledge‐driven deep learning for fast MR imaging: Undersampled MR image reconstruction from supervised to un‐supervised learning

Contact Info

Product

Resources

About