Abstract:Background: Compressed sensing has been employed to accelerate magnetic resonance imaging by sampling fewer measurements. However, conventional iterative optimization-based CS-MRI are time-consuming for iterative calculations and often share poor generalization ability on multicontrast datasets. Most deep-learning-based CS-MRI focus on learning an end-to-end mapping while ignoring some prior knowledge existed in MR images. Purpose: We propose an iterative fusion model to integrate the image and gradient-based priors into reconstruction via convolutional neural network models while maintaining high quality and preserving the detailed information as well. Methods: We propose a gradient-enhanced fusion network (GFN) for fast and accurate MRI reconstruction, in which dense blocks with dilated convolution and dense residual learnings are used to capture abundant features with fewer parameters. Meanwhile, decomposed gradient maps containing obvious structural information are introduced into the network to enhance the reconstructed images. Besides this, gradient-based priors along directions X and Y are exploited to learn adaptive tight frames for reconstructing the desired image contrast and edges by respective gradient fusion networks. After that, both image and gradient priors are fused in the proposed optimization model, in which we employ the l2 -norm to promote the sparsity of gradient priors. The proposed fusion model can effectively help to capture edge structures in the gradient images and to preserve more detailed information of MR images. Results: Experimental results demonstrate that the proposed method outperforms several CS-MRI methods in terms of peak signal-to-noise (PSNR), the structural similarity index (SSIM), and visualizations on three sampling masks with different rates. Noteworthy, to evaluate the generalization ability, the proposed model conducts cross-center training and testing experiments for all three modalities and shares more stable performance compared than other approaches. In addition, the proposed fusion model is applied to other comparable deep learning methods. The quantitative results show that the reconstruction results of these methods are obviously improved. Conclusions: The gradient-based priors reconstructed from GFNs can effectively enhance the edges and details of under-sampled data. The proposed fusion model integrates image and gradient priors using l2 -norm can effectively improve the generalization ability on multicontrast datasets reconstruction.

A Very Deep Densely Connected Network for Compressed Sensing MRI

EiCSNet: Efficient Iterative Neural Network for Compressed Sensing Reconstruction

A Deep Ensemble Network for Compressed Sensing MRI

MS-DRDNet: Optimization-Inspired Deep Compressive Sensing Network for MRI

Compressed Sensing MR Image Reconstruction Via a Deep Frequency-Division Network.

Deep Convolutional Neural Network for Compressive Sensing of Magnetic Resonance Images

Cascaded Dilated Dense Network with Two-step Data Consistency for MRI Reconstruction

Compressed Sensing MRI via a Multi-scale Dilated Residual Convolution Network

ADMM-Net: A Deep Learning Approach for Compressive Sensing MRI

A Deep Error Correction Network for Compressed Sensing MRI

Compressed Sensing MRI Using a Recursive Dilated Network.

Triple-D Network for Efficient Undersampled Magnetic Resonance Images Reconstruction.

MD-Recon-Net: A Parallel Dual-Domain Convolutional Neural Network for Compressed Sensing MRI

A Segmentation-aware Deep Fusion Network for Compressed Sensing MRI

A Deep Information Sharing Network for Multi-contrast Compressed Sensing MRI Reconstruction

Attention-Enhanced Convolutional Iterative Unrolling Network for Under-Sampled MRI Reconstruction

A Dictionary Learning Based Unsupervised Neural Network for Single Image Compressed Sensing

Deep ADMM-Net for Compressive Sensing MRI

Deep compressed sensing MRI via a gradient-enhanced fusion model

Deep Geometric Distillation Network for Compressive Sensing MRI