Abstract:Inspired by recent advances in deep learning, we propose a framework for reconstructing dynamic sequences of 2D cardiac magnetic resonance (MR) images from undersampled data using a deep cascade of convolutional neural networks (CNNs) to accelerate the data acquisition process. In particular, we address the case where data is acquired using aggressive Cartesian undersampling. Firstly, we show that when each 2D image frame is reconstructed independently, the proposed method outperforms state-of-the-art 2D compressed sensing approaches such as dictionary learning-based MR image reconstruction, in terms of reconstruction error and reconstruction speed. Secondly, when reconstructing the frames of the sequences jointly, we demonstrate that CNNs can learn spatio-temporal correlations efficiently by combining convolution and data sharing approaches. We show that the proposed method consistently outperforms state-of-the-art methods and is capable of preserving anatomical structure more faithfully up to 11-fold undersampling. Moreover, reconstruction is very fast: each complete dynamic sequence can be reconstructed in less than 10s and, for the 2D case, each image frame can be reconstructed in 23ms, enabling real-time applications.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to rapidly and high - quality reconstruct image sequences from undersampled data in Dynamic Cardiac MRI. Specifically, the authors propose a method using a deep convolutional neural network (CNNs) cascading framework to accelerate the data acquisition process and still maintain the fidelity of anatomical structures under high - magnification undersampling. ### Main Problems 1. **Slow Data Acquisition**: Traditional magnetic resonance imaging (MRI) has high requirements for patients due to its inherently slow data acquisition process, making this imaging method costly and difficult to obtain. 2. **Artifacts Caused by Undersampling**: In order to accelerate data acquisition, the method of undersampling can be adopted, but this will lead to a violation of the Nyquist - Shannon sampling theorem and produce aliasing artifacts. 3. **High - Quality Image Reconstruction**: In the case of undersampling, how to efficiently reconstruct high - quality images, especially in dynamic imaging, requires considering spatio - temporal correlations. ### Solutions The authors propose a deep - learning - based method, especially using a deep convolutional neural network (CNNs) cascading framework to solve the above problems. The specific contributions are as follows: 1. **Independent Frame Reconstruction**: When each 2D image frame is reconstructed independently, the proposed method is superior to existing 2D compressed sensing methods, such as MRI image reconstruction based on dictionary learning, in terms of reconstruction error and reconstruction speed. 2. **Joint Frame Reconstruction**: When jointly reconstructing the frames of an image sequence, CNNs can effectively learn spatio - temporal correlations by combining convolution and data - sharing methods. 3. **Performance under High - Magnification Undersampling**: Experimental results show that this method can still maintain relatively high image quality under an undersampling rate of up to 11 times. 4. **Efficient Reconstruction Speed**: Each complete dynamic sequence can be reconstructed within less than 10 seconds, and for the 2D case, each image frame can be reconstructed within 23 milliseconds, which is suitable for real - time applications. ### Technical Details 1. **Data Consistency Layer**: A data consistency layer is introduced to ensure that the network output is consistent with the known k - space data, thereby improving the reconstruction quality. 2. **Data Sharing Layer**: Utilize the spatio - temporal correlations between adjacent frames, and fill in the missing k - space samples through the data - sharing method to further reduce aliasing artifacts. 3. **Cascading Network Architecture**: A cascading CNN architecture is constructed, and through multiple iterations, de - aliasing and data consistency correction are carried out to improve the reconstruction effect. ### Experimental Results The authors conducted experiments on 10 fully - sampled short - axis cardiac cine - MRI scan data sets to verify the effectiveness of the proposed method. The experimental results show that this method can significantly outperform existing compressed - sensing methods at different undersampling rates, especially outstanding at high - magnification undersampling. In conclusion, this paper proposes an innovative deep - learning method that can achieve rapid and high - quality image reconstruction in dynamic cardiac MRI and has important clinical application value.

A Deep Cascade of Convolutional Neural Networks for Dynamic MR Image Reconstruction

A Deep Cascade of Convolutional Neural Networks for Dynamic MR Image Reconstruction.

A Deep Cascade of Convolutional Neural Networks for MR Image Reconstruction

Deep Learning for Under-Sampled Non-Cartesian ASL MRI Reconstruction

Parallel Imaging and Convolutional Neural Network Combined Fast MR Image Reconstruction: Applications in Low-Latency Accelerated Real-Time Imaging.

Deep Cardiac MRI Reconstruction with ADMM

Deep Network Cascade for Dynamic Cardiac MRI Reconstruction With Motion Feature Incorporation and the Fourier Neural Attention

Data-Consistent Non-Cartesian Deep Subspace Learning for Efficient Dynamic MR Image Reconstruction

Deep learning within a priori temporal feature spaces for large-scale dynamic MR image reconstruction: Application to 5-D cardiac MR Multitasking

Dynamic MR imaging reconstruction by three-dimensional dictionary learning

Stochastic Deep Compressive Sensing for the Reconstruction of Diffusion Tensor Cardiac MRI

Fully Unsupervised Dynamic MRI Reconstruction via Diffeo-Temporal Equivariance

Unsupervised reconstruction of accelerated cardiac cine MRI using Neural Fields

Dynamic MRI of the Abdomen Using Parallel Non-Cartesian Convolutional Recurrent Neural Networks.

A Deep Learning Framework for Cardiac MR Under-Sampled Image Reconstruction with a Hybrid Spatial and k-Space Loss Function

Real-time Cardiovascular MR with Spatio-temporal Artifact Suppression using Deep Learning - Proof of Concept in Congenital Heart Disease

End-to-end Adaptive Dynamic Subsampling and Reconstruction for Cardiac MRI

SelfCoLearn: Self-supervised Collaborative Learning for Accelerating Dynamic MR Imaging.

A Hybrid Frequency-domain/Image-domain Deep Network for Magnetic Resonance Image Reconstruction

Joint reconstruction framework of compressed sensing and nonlinear parallel imaging for dynamic cardiac magnetic resonance imaging

DIR3D: Cascaded Dual-Domain Inter-Scale Mutual Reinforcement 3D Network for Highly Accelerated 3D MR Image Reconstruction