A densely interconnected network for deep learning accelerated MRI

Ottesen, Jon André; Caan, Matthan W.A.; Groote, Inge Rasmus; Bjørnerud, Atle

doi:10.1007/s10334-022-01041-3

Cited by 9 publications

(4 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…DEMO can be flexibly incorporated into both model-based and unrolled deep neural network CS-MRI methods since it is independent of any backbone algorithm. Ottesen et al [ 85 ] implemented the Densely Interconnected Residual Cascading Network (DIRCN) for MRI reconstruction, drawing inspiration from the end-to-end variational network. The method utilized input-level connections and long-range skip connections to enhance MRI quality at high acceleration rates.…”

Section: Papers Improving Deep Mri Reconstruction Methodsmentioning

confidence: 99%

Emerging Trends in Fast MRI Using Deep-Learning Reconstruction on Undersampled k-Space Data: A Systematic Review

Singh,

Monga,

de Moura

et al. 2023

Bioengineering

View full text Add to dashboard Cite

Magnetic Resonance Imaging (MRI) is an essential medical imaging modality that provides excellent soft-tissue contrast and high-resolution images of the human body, allowing us to understand detailed information on morphology, structural integrity, and physiologic processes. However, MRI exams usually require lengthy acquisition times. Methods such as parallel MRI and Compressive Sensing (CS) have significantly reduced the MRI acquisition time by acquiring less data through undersampling k-space. The state-of-the-art of fast MRI has recently been redefined by integrating Deep Learning (DL) models with these undersampled approaches. This Systematic Literature Review (SLR) comprehensively analyzes deep MRI reconstruction models, emphasizing the key elements of recently proposed methods and highlighting their strengths and weaknesses. This SLR involves searching and selecting relevant studies from various databases, including Web of Science and Scopus, followed by a rigorous screening and data extraction process using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. It focuses on various techniques, such as residual learning, image representation using encoders and decoders, data-consistency layers, unrolled networks, learned activations, attention modules, plug-and-play priors, diffusion models, and Bayesian methods. This SLR also discusses the use of loss functions and training with adversarial networks to enhance deep MRI reconstruction methods. Moreover, we explore various MRI reconstruction applications, including non-Cartesian reconstruction, super-resolution, dynamic MRI, joint learning of reconstruction with coil sensitivity and sampling, quantitative mapping, and MR fingerprinting. This paper also addresses research questions, provides insights for future directions, and emphasizes robust generalization and artifact handling. Therefore, this SLR serves as a valuable resource for advancing fast MRI, guiding research and development efforts of MRI reconstruction for better image quality and faster data acquisition.

show abstract

Section: Papers Improving Deep Mri Reconstruction Methodsmentioning

confidence: 99%

Emerging Trends in Fast MRI Using Deep-Learning Reconstruction on Undersampled k-Space Data: A Systematic Review

Singh,

Monga,

de Moura

et al. 2023

Bioengineering

View full text Add to dashboard Cite

show abstract

“…Ottesen et al [2] proposed a densely connected cascading deep learning reconstruction framework to improve accelerated MRI reconstruction. The authors modified a cascading deep learning reconstruction framework by incorporating three architectural modifications, namely input-level dense connections, an improved deep learning sub-network, and long-range skip-connections.…”

Section: Literature Reviewmentioning

confidence: 99%

Super-Resolution of Brain MRI via U-Net Architecture

Aryan¹

2023

IJACSA

View full text Add to dashboard Cite

This paper proposes a U-Net-based deep learning architecture for the task of super-resolution of lower resolution brain magnetic resonance images (MRI). The proposed system, called MRI-Net, is designed to learn the mapping between lowresolution and high-resolution MRI images. The system is trained using 50-800 2D MRI scans, depending on the architecture, and is evaluated using peak signal-to-noise ratio (PSNR) metrics on 10 randomly selected images. The proposed U-Net architecture outperforms current state-of-the-art networks in terms of PSNR when evaluated with a 3 x 3 resolution downsampling index. The system's ability to super-resolve MRI scans has the potential to enable physicians to detect pathologies better and perform a wider range of applications. The symmetrical downsampling pipeline used in this study allows for generically representing low-resolution MRI scans to highlight proof of concept for the U-Net-based approach. The system is implemented on PyTorch 1.9.0 with NVIDIA GPU processing to speed up training time. U-Net is a promising tool for medical applications in MRI, which can provide accurate and highquality images for better diagnoses and treatment plans. The proposed approach has the potential to reduce the costs associated with high-resolution MRI scans by providing a solution for enhancing the image quality of low-resolution scans.

show abstract

“…Densely connected networks [48] maximize the capability of the network by reusing features. The input of the succeeding layers is more varied and more effective when feature maps from various layers are combined.…”

Section: Dense Blockmentioning

confidence: 99%

De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates

et al. 2022

View full text Add to dashboard Cite

When sparsely sampled data are used to accelerate magnetic resonance imaging (MRI), conventional reconstruction approaches produce significant artifacts that obscure the content of the image. To remove aliasing artifacts, we propose an advanced convolutional neural network (CNN) called fully dense attention CNN (FDA-CNN). We updated the Unet model with the fully dense connectivity and attention mechanism for MRI reconstruction. The main benefit of FDA-CNN is that an attention gate in each decoder layer increases the learning process by focusing on the relevant image features and provides a better generalization of the network by reducing irrelevant activations. Moreover, densely interconnected convolutional layers reuse the feature maps and prevent the vanishing gradient problem. Additionally, we also implement a new, proficient under-sampling pattern in the phase direction that takes low and high frequencies from the k-space both randomly and non-randomly. The performance of FDA-CNN was evaluated quantitatively and qualitatively with three different sub-sampling masks and datasets. Compared with five current deep learning-based and two compressed sensing MRI reconstruction techniques, the proposed method performed better as it reconstructed smoother and brighter images. Furthermore, FDA-CNN improved the mean PSNR by 2 dB, SSIM by 0.35, and VIFP by 0.37 compared with Unet for the acceleration factor of 5.

show abstract

A densely interconnected network for deep learning accelerated MRI

Cited by 9 publications

References 48 publications

Emerging Trends in Fast MRI Using Deep-Learning Reconstruction on Undersampled k-Space Data: A Systematic Review

Emerging Trends in Fast MRI Using Deep-Learning Reconstruction on Undersampled k-Space Data: A Systematic Review

Super-Resolution of Brain MRI via U-Net Architecture

De-Aliasing and Accelerated Sparse Magnetic Resonance Image Reconstruction Using Fully Dense CNN with Attention Gates

Contact Info

Product

Resources

About