CS-MRI Reconstruction Using an Improved GAN with Dilated Residual Networks and Channel Attention Mechanism.

Compressed sensing (CS) MRI has shown great potential in enhancing time efficiency. Deep learning techniques, specifically generative adversarial networks (GANs), have emerged as potent tools for speedy CS-MRI reconstruction. Yet, as the complexity of deep learning reconstruction models increases, this can lead to prolonged reconstruction time and challenges in achieving convergence. In this study, we present a novel GAN-based model that delivers superior performance without the model complexity escalating. Our generator module, built on the U-net architecture, incorporates dilated residual (DR) networks, thus expanding the network's receptive field without increasing parameters or computational load. At every step of the downsampling path, this revamped generator module includes a DR network, with the dilation rates adjusted according to the depth of the network layer. Moreover, we have introduced a channel attention mechanism (CAM) to distinguish between channels and reduce background noise, thereby focusing on key information. This mechanism adeptly combines global maximum and average pooling approaches to refine channel attention. We conducted comprehensive experiments with the designed model using public domain MRI datasets of the human brain. Ablation studies affirmed the efficacy of the modified modules within the network. Incorporating DR networks and CAM elevated the peak signal-to-noise ratios (PSNR) of the reconstructed images by about 1.2 and 0.8 dB, respectively, on average, even at 10× CS acceleration. Compared to other relevant models, our proposed model exhibits exceptional performance, achieving not only excellent stability but also outperforming most of the compared networks in terms of PSNR and SSIM. When compared with U-net, DR-CAM-GAN's average gains in SSIM and PSNR were 14% and 15%, respectively. Its MSE was reduced by a factor that ranged from two to seven. The model presents a promising pathway for enhancing the efficiency and quality of CS-MRI reconstruction.

Deep compressed sensing MRI via a gradient-enhanced fusion model.

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.

Undersampling patterns in k-space for compressed sensing MRI using two-dimensional Cartesian sampling.

In compressed sensing magnetic resonance imaging (CS-MRI), undersampling of k-space is performed to achieve faster imaging. For this process, it is important to acquire data randomly, and an optimal random undersampling pattern is required. However, random undersampling is difficult in two-dimensional (2D) Cartesian sampling. In this study, the effect of random undersampling patterns on image reconstruction was clarified using phantom and in vivo MRI, and a sampling pattern relevant for 2D Cartesian sampling in CS-MRI is suggested. The precision of image restoration was estimated with various acceleration factors and extents for the fully sampled central region of k-space. The root-mean-square error, structural similarity index, and modulation transfer function were measured, and visual assessments were also performed. The undersampling pattern was shown to influence the precision of image restoration, and an optimal undersampling pattern should be used to improve image quality; therefore, we suggest that the ideal undersampling pattern in CS-MRI for 2D Cartesian sampling is one with a high extent for the fully sampled central region of k-space.

Compressed sensing MRI via fast linearized preconditioned alternating direction method of multipliers.

BACKGROUND: The challenge of reconstructing a sparse medical magnetic resonance image based on compressed sensing from undersampled k-space data has been investigated within recent years. As total variation (TV) performs well in preserving edge, one type of approach considers TV-regularization as a sparse structure to solve a convex optimization problem. Nevertheless, this convex optimization problem is both nonlinear and nonsmooth, and thus difficult to handle, especially for a large-scale problem. Therefore, it is essential to develop efficient algorithms to solve a very broad class of TV-regularized problems. METHODS: In this paper, we propose an efficient algorithm referred to as the fast linearized preconditioned alternating direction method of multipliers (FLPADMM), to solve an augmented TV-regularized model that adds a quadratic term to enforce image smoothness. Because of the separable structure of this model, FLPADMM decomposes the convex problem into two subproblems. Each subproblem can be alternatively minimized by augmented Lagrangian function. Furthermore, a linearized strategy and multistep weighted scheme can be easily combined for more effective image recovery. RESULTS: The method of the present study showed improved accuracy and efficiency, in comparison to other methods. Furthermore, the experiments conducted on in vivo data showed that our algorithm achieved a higher signal-to-noise ratio (SNR), lower relative error (Rel.Err), and better structural similarity (SSIM) index in comparison to other state-of-the-art algorithms. CONCLUSIONS: Extensive experiments demonstrate that the proposed algorithm exhibits superior performance in accuracy and efficiency than conventional compressed sensing MRI algorithms.

Diagnostic quality assessment of compressed sensing accelerated magnetic resonance neuroimaging.

PURPOSE: To determine the efficacy of compressed sensing (CS) reconstructions for specific clinical magnetic resonance neuroimaging applications beyond more conventional acceleration techniques such as parallel imaging (PI) and low-resolution acquisitions. MATERIALS AND METHODS: Raw k-space data were acquired from five healthy volunteers on a 3T scanner using a 32-channel head coil using T2 -FLAIR, FIESTA-C, time of flight (TOF), and spoiled gradient echo (SPGR) sequences. In a series of blinded studies, three radiologists independently evaluated CS, PI (GRAPPA), and low-resolution images at up to 5× accelerations. Synthetic T2 -FLAIR images with artificial lesions were used to assess diagnostic accuracy for CS reconstructions. RESULTS: CS reconstructions were of diagnostically acceptable quality at up to 4× acceleration for T2 -FLAIR and FIESTA-C (average qualitative scores 3.7 and 4.3, respectively, on a 5-point scale at 4× acceleration), and at up to 3× acceleration for TOF and SPGR (average scores 4.0 and 3.7, respectively, at 3× acceleration). The qualitative scores for CS reconstructions were significantly better than low-resolution images for T2 -FLAIR, FIESTA-C, and TOF and significantly better than GRAPPA for TOF and SPGR (Wilcoxon signed rank test, P < 0.05) with no significant difference found otherwise. Diagnostic accuracy was acceptable for both CS and low-resolution images at up to 3× acceleration (area under the ROC curve 0.97 and 0.96, respectively.) CONCLUSION: Mild to moderate accelerations are possible for those sequences by a combined CS and PI reconstruction. Nevertheless, for certain sequences/applications one might mildly reduce the acquisition time by appropriately reducing the imaging resolution rather than the more complicated CS reconstruction. J. Magn. Reson. Imaging 2016;44:433-444.

Real time dynamic MRI by exploiting spatial and temporal sparsity.

Online imaging requires that the reconstruction of current frame only depends on the previous frames, and real time imaging is the desired case. In this work, we propose a novel scheme for real time dynamic magnetic resonance imaging (dMRI) reconstruction. Different from previous methods, the reconstructions of the second frame to the last frame are independent in our scheme, which only require the first frame as the reference image. Therefore, this scheme can be naturally implemented in parallel. After the first frame is reconstructed, all the later frames can be processed as soon as the k-space data are acquired. As an extension of the conventional spatial total variation, a new online model called dynamic total variation is used to exploit the sparsity on both spatial and temporal domains in dMRI. In real time dMRI, each frame is required to be reconstructed very fast. We then design a novel reweighted least squares algorithm to solve the challenging problem. Motivated by the special structure of partial Fourier transform in sparse MRI, this algorithm is accelerated by the preconditioned conjugate gradient descent method. The proposed method is compared with 4 state-of-the-art online and offline methods on two in-vivo cardiac dMRI datasets. The experimental results show that our method significantly outperforms previous online methods, and is comparable to the offline methods in terms of reconstruction accuracy.

Fast multi-contrast MRI reconstruction.

Multi-contrast magnetic resonance imaging (MRI) is a useful technique to aid clinical diagnosis. This paper proposes an efficient algorithm to jointly reconstruct multiple T1/T2-weighted images of the same anatomical cross section from partially sampled k-space data. The joint reconstruction problem is formulated as minimizing a linear combination of three terms, corresponding to a least squares data fitting, joint total variation (TV) and group wavelet-sparsity regularization. It is rooted in two observations: 1) the variance of image gradients should be similar for the same spatial position across multiple contrasts; 2) the wavelet coefficients of all images from the same anatomical cross section should have similar sparse modes. To efficiently solve this problem, we decompose it into joint TV regularization and group sparsity subproblems, respectively. Finally, the reconstructed image is obtained from the weighted average of solutions from the two subproblems, in an iterative framework. Experiments demonstrate the efficiency and effectiveness of the proposed method compared to existing multi-contrast MRI methods.

Exploiting the wavelet structure in compressed sensing MRI.

Sparsity has been widely utilized in magnetic resonance imaging (MRI) to reduce k-space sampling. According to structured sparsity theories, fewer measurements are required for tree sparse data than the data only with standard sparsity. Intuitively, more accurate image reconstruction can be achieved with the same number of measurements by exploiting the wavelet tree structure in MRI. A novel algorithm is proposed in this article to reconstruct MR images from undersampled k-space data. In contrast to conventional compressed sensing MRI (CS-MRI) that only relies on the sparsity of MR images in wavelet or gradient domain, we exploit the wavelet tree structure to improve CS-MRI. This tree-based CS-MRI problem is decomposed into three simpler subproblems then each of the subproblems can be efficiently solved by an iterative scheme. Simulations and in vivo experiments demonstrate the significant improvement of the proposed method compared to conventional CS-MRI algorithms, and the feasibleness on MR data compared to existing tree-based imaging algorithms.

Enhancement of the low resolution image quality using randomly sampled data for multi-slice MR imaging.

Low resolution images are often acquired in in vivo MR applications involving in large field-of-view (FOV) and high speed imaging, such as, whole-body MRI screening and functional MRI applications. In this work, we investigate a multi-slice imaging strategy for acquiring low resolution images by using compressed sensing (CS) MRI to enhance the image quality without increasing the acquisition time. In this strategy, low resolution images of all the slices are acquired using multiple-slice imaging sequence. In addition, extra randomly sampled data in one center slice are acquired by using the CS strategy. These additional randomly sampled data are multiplied by the weighting functions generated from low resolution full k-space images of the two slices, and then interpolated into the k-space of other slices. In vivo MR images of human brain were employed to investigate the feasibility and the performance of the proposed method. Quantitative comparison between the conventional low resolution images and those from the proposed method was also performed to demonstrate the advantage of the method.

Improving multi-channel compressed sensing MRI with reweighted l 1 minimization.

Integrating compressed sensing (CS) and parallel imaging (PI) with multi-channel receiver has proven to be an effective technology to speed up magnetic resonance imaging (MRI). In this paper, we propose a method that extends the reweighted l 1 minimization to the CS-MRI with multi-channel data. The method applies a reweighted l 1 minimization algorithm to reconstruct each channel image, and then generates the final image by a sum-of-squares method. Computer simulations based on synthetic data and in vivo MRI imaging data show that the new method can improve the reconstruction quality at a slightly increased computation cost.