TRANS-Net: Transformer-Enhanced Residual-Error AlterNative Suppression Network for MRI Reconstruction

Hu, Dianlin; Zhang, Yikun; Zhu, Jianfeng; Liu, Qiegen; Chen, Yang

doi:10.1109/tim.2022.3205684

Cited by 8 publications

(7 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The unrolling methods learn optimal, robust parameters, or other complex prior knowledges from big datasets to achieve the best performances for iterative MRI reconstruction models. 1,4,[141][142][143][144][145][146][147][148][149][150][151][152][153][154] The data-driven methods trained the model offline on large datasets to extract valuable knowledge such as the regularizers, mapping and denoising, and then used these learned knowledge for online inferring to reconstruct high-quality images from undersampled k-space data. 2,3,50,120,155-162 Different network architectures have been investigated 86, and uncertainty has also been estimated and introduced to improve the reconstruction performance.…”

Section: Supervised DL For Fast Mrimentioning

confidence: 99%

“…Supervised learning‐based methods have two main categories, model unrolling‐based methods and data‐driven methods. The unrolling methods learn optimal, robust parameters, or other complex prior knowledges from big datasets to achieve the best performances for iterative MRI reconstruction models 1,4,141–154 . The data‐driven methods trained the model offline on large datasets to extract valuable knowledge such as the regularizers, mapping and denoising, and then used these learned knowledge for online inferring to reconstruct high‐quality images from undersampled k‐space data 2,3,50,120,155–162 .…”

Section: Paradigm Shift and Applications For Mri Reconstructionmentioning

confidence: 99%

See 1 more Smart Citation

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

Self Cite

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

Section: Supervised DL For Fast Mrimentioning

confidence: 99%

Section: Paradigm Shift and Applications For Mri Reconstructionmentioning

confidence: 99%

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

Self Cite

View full text Add to dashboard Cite

show abstract

“…We compare the de-artifact results of this paper with the other 6 representative results. The comparison methods include DeblurGAN (Kupyn et al 2018), TGVNN (Wang et al 2020), CRNN (Qin et al 2019), RGAN (Lyu et al 2021), Trans-Net (Hu et al 2022), and RMT-GAN (Lyu et al 2023). Among them, DeblurGAN is the original de-artifact method based on GAN.…”

Section: Comparative Experimentsmentioning

confidence: 99%

“…The new networks enhance the performance in MRI de-artifact work. Hu et al (2022) and Wang et al (2024) combine the improved transformer networks with other network structures to restore blur details in images. However, the existing MRI artifact reduction methods pay little attention to the texture edges of blurred images.…”

Section: Introductionmentioning

confidence: 99%

A cardiac MRI motion artifact reduction method based on edge enhancement network

Jiang,

Zhang,

et al. 2024

Phys. Med. Biol.

View full text Add to dashboard Cite

Cardiac magnetic resonance imaging (MRI) usually requires a long acquisition time. The movement of the patients during MRI acquisition will produce image artifacts. Previous studies have shown that clear MR image texture edges are of great significance for pathological diagnosis. In this paper, a motion artifact reduction method for cardiac MRI based on edge enhancement network is proposed. Firstly, the four-plane normal vector adaptive fractional differential mask is applied to extract the edge features of blurred images. The four-plane normal vector method can reduce the noise information in the edge feature maps. The adaptive fractional order is selected according to the normal mean gradient and the local Gaussian curvature entropy of the images. Then, the extracted edge feature maps and blurred images are input into the de-artifact network. In this network, the edge fusion feature extraction network and the edge fusion Transformer network are specially designed. The former combines the edge feature maps with the fuzzy feature maps to extract the edge feature information. The latter combines the edge attention network and the fuzzy attention network, which can focus on the blurred image edges. Finally, extensive experiments show that the proposed method can obtain higher Peak Signal-to-Noise Ratio(PSNR) and Structural Similarity Index Measure(SSIM) compared to state-of-art methods. The de-artifact images have clear texture edges.

show abstract

“…For DL-based reconstruction, the proposed reconstruction models can be categorized into data-driven and model-driven unrolled networks. Part of the data-driven model was designed in an end-to-end manner to fit the mapping between the input measurements and reconstruction targets [8]- [10]. However, these methods still present problems such as lack of interpretability, and a relatively large dataset is required to fit the mapping.…”

Section: Introductionmentioning

confidence: 99%

MEDL‐Net: A model‐based neural network for MRI reconstruction with enhanced deep learned regularizers

Qiao

Huang

2023

Magnetic Resonance in Med

View full text Add to dashboard Cite

Purpose: To improve the MRI reconstruction performance of model-based networks and to alleviate their large demand for GPU memory.Methods: A model-based neural network with enhanced deep learned regularizers (MEDL-Net) was proposed. The MEDL-Net is separated into several submodules, each of which consists of several cascades to mimic the optimization steps in conventional MRI reconstruction algorithms. Information from shallow cascades is densely connected to latter ones to enrich their inputs in each submodule, and additional revising blocks (RB) are stacked at the end of the submodules to bring more flexibility. Moreover, a composition loss function was designed to explicitly supervise RBs.Results: Network performance was evaluated on a publicly available dataset.The MEDL-Net quantitatively outperforms the state-of-the-art methods on different MR image sequences with different acceleration rates (four-fold and six-fold). Moreover, the reconstructed images showed that the detailed textures are better preserved. In addition, fewer cascades are required when achieving the same reconstruction results compared with other model-based networks. Conclusion:In this study, a more efficient model-based deep network was proposed to reconstruct MR images. The experimental results indicate that the proposed method improves reconstruction performance with fewer cascades, which alleviates the large demand for GPU memory.

show abstract

TRANS-Net: Transformer-Enhanced Residual-Error AlterNative Suppression Network for MRI Reconstruction

Cited by 8 publications

References 50 publications

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

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

A cardiac MRI motion artifact reduction method based on edge enhancement network

MEDL‐Net: A model‐based neural network for MRI reconstruction with enhanced deep learned regularizers

Contact Info

Product

Resources

About