Enhanced Deep-Learning-Based Magnetic Resonance Image Reconstruction by Leveraging Prior Subject-Specific Brain Imaging: Proof-of-Concept Using a Cohort of Presumed Normal Subjects

Souza, Roberto; Beauferris, Youssef; Loos, Wallace; Frayne, Richard

doi:10.1109/jstsp.2020.3001525

Cited by 15 publications

(8 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In 41.3% of the reviewed studies, parallel imaging is combined with CS to exploit the k-space signal redundancy collected by multiple receiver coils. Similar to multicontrast reconstruction, for separate imaging coils, many studies use separate input and outputs channels [17], [40], [57], [70], [75], [76], [78], [80], [93]- [95]. The reconstructed images for each coil are then combined by the sum-of-squares.…”

Section: Parallel Imaging With Coil Redundancymentioning

confidence: 99%

AI-Based Reconstruction for Fast MRI—A Systematic Review and Meta-Analysis

et al. 2022

View full text Add to dashboard Cite

show abstract

Section: Parallel Imaging With Coil Redundancymentioning

confidence: 99%

AI-Based Reconstruction for Fast MRI—A Systematic Review and Meta-Analysis

et al. 2022

View full text Add to dashboard Cite

show abstract

“…For example, a fusion network was developed to learn the latent representation of multi-modal MR images on the synthesis task [27]. MR images scanned from previous visits for the same patient can be used as reference to accelerate the acquisition in the follow-up visits [28]. For accelerated multi-modal MRI, methods have been proposed to use MR images of one modality as reference to assist the reconstruction of under-sampled MR images of another modality.…”

Section: A Multi-modal Mri Reconstructionmentioning

confidence: 99%

Fast T2w/FLAIR MRI Acquisition by Optimal Sampling of Information Complementary to Pre-acquired T1w MRI

Yang¹,

Li²,

Liu³

et al. 2021

Preprint

View full text Add to dashboard Cite

Recent studies on T1-assisted MRI reconstruction for under-sampled images of other modalities have demonstrated the potential of further accelerating MRI acquisition of other modalities. Most of the state-ofthe-art approaches have achieved improvement through the development of network architectures for fixed undersampling patterns, without fully exploiting the complementary information between modalities. Although existing under-sampling pattern learning algorithms can be simply modified to allow the fully-sampled T1-weighted MR image to assist the pattern learning, no significant improvement on the reconstruction task can be achieved. To this end, we propose an iterative framework to optimize the undersampling pattern for MRI acquisition of another modality that can complement the fully-sampled T1-weighted MR image at different under-sampling factors, while jointly optimizing the T1-assisted MRI reconstruction model. Specifically, our proposed method exploits the difference of latent information between the two modalities for determining the sampling patterns that can maximize the assistance power of T1-weighted MR image in improving the MRI reconstruction. We have demonstrated superior performance of our learned under-sampling patterns on a public dataset, compared to commonly used under-sampling patterns and state-of-the-art methods that can jointly optimize both the reconstruction network and the under-sampling pattern, up to 8-fold under-sampling factor.

show abstract

“…With physiology-related redundancy between MR images, reconstructing under-sampled MRI with auxiliary information from reference images is a promising way. Regarding multiple visits of the same patient, Souza et al [19] proposed to accelerate MRI follow-ups with early intra-subject images. Weizman et al [20], [21] proposed a more flexible solution in that they also support the case when the reference is not fully available.…”

Section: B Reference-based Mri Reconstructionmentioning

confidence: 99%

Multi-Modal MRI Reconstruction Assisted with Spatial Alignment Network

Xuan,

Xiang,

Huang

et al. 2021

Preprint

View full text Add to dashboard Cite

In clinical practice, magnetic resonance imaging (MRI) with multiple contrasts is usually acquired in a single study to assess different properties of the same region of interest in human body. The whole acquisition process can be accelerated by having one or more modalities under-sampled in the k-space. Recent researches demonstrate that, considering the redundancy between different contrasts or modalities, a target MRI modality under-sampled in the k-space can be more efficiently reconstructed with a fully-sampled MRI contrast as the reference modality. However, we find that the performance of the above multi-modal reconstruction can be negatively affected by subtle spatial misalignment between different contrasts, which is actually common in clinical practice. In this paper, to compensate for such spatial misalignment, we integrate the spatial alignment network with multi-modal reconstruction towards better reconstruction quality of the target modality. First, the spatial alignment network estimates the spatial misalignment between the fully-sampled reference and the under-sampled target images, and warps the reference image accordingly. Then, the aligned fully-sampled reference image joins the multi-modal reconstruction of the under-sampled target image. Also, considering the contrast difference between the target and the reference images, we particularly design the cross-modality-synthesisbased registration loss, in combination with the reconstruction loss, to jointly train the spatial alignment network and the reconstruction network. Experiments on both clinical MRI and multi-coil k-space raw data demonstrate the superiority and robustness of multi-modal MRI reconstruction empowered with our spatial alignment network. Our code is publicly available at https://github.com/woxuankai/SpatialAlignmentNetwork.

show abstract

Enhanced Deep-Learning-Based Magnetic Resonance Image Reconstruction by Leveraging Prior Subject-Specific Brain Imaging: Proof-of-Concept Using a Cohort of Presumed Normal Subjects

Cited by 15 publications

References 35 publications

AI-Based Reconstruction for Fast MRI—A Systematic Review and Meta-Analysis

AI-Based Reconstruction for Fast MRI—A Systematic Review and Meta-Analysis

Fast T2w/FLAIR MRI Acquisition by Optimal Sampling of Information Complementary to Pre-acquired T1w MRI

Multi-Modal MRI Reconstruction Assisted with Spatial Alignment Network

Contact Info

Product

Resources

About