Compressed sensing MR image reconstruction using a motion-compensated reference

Reference-Driven Undersampled MR Image Reconstruction Using Wavelet Sparsity-Constrained Deep Image Prior

Deep learning has shown potential in significantly improving performance for undersampled magnetic resonance (MR) image reconstruction. However, one challenge for the application of deep learning to clinical scenarios is the requirement of large, high-quality patient-based datasets for network training. In this paper, we propose a novel deep learning-based method for undersampled MR image reconstruction that does not require pre-training procedure and pre-training datasets. The proposed reference-driven method using wavelet sparsity-constrained deep image prior (RWS-DIP) is based on the DIP framework and thereby reduces the dependence on datasets. Moreover, RWS-DIP explores and introduces structure and sparsity priors into network learning to improve the efficiency of learning. By employing a high-resolution reference image as the network input, RWS-DIP incorporates structural information into network. RWS-DIP also uses the wavelet sparsity to further enrich the implicit regularization of traditional DIP by formulating the training of network parameters as a constrained optimization problem, which is solved using the alternating direction method of multipliers (ADMM) algorithm. Experiments on in vivo MR scans have demonstrated that the RWS-DIP method can reconstruct MR images more accurately and preserve features and textures from undersampled k-space measurements.

Coupled Dictionary Learning for Multi-Contrast MRI Reconstruction

Magnetic resonance (MR) imaging tasks often involve multiple contrasts, such as T1-weighted, T2-weighted and fluid-attenuated inversion recovery (FLAIR) data. These contrasts capture information associated with the same underlying anatomy and thus exhibit similarities in either structure level or gray level. In this paper, we propose a coupled dictionary learning based multi-contrast MRI reconstruction (CDLMRI) approach to leverage the dependency correlation between different contrasts for guided or joint reconstruction from their under-sampled k -space data. Our approach iterates between three stages: coupled dictionary learning, coupled sparse denoising, and enforcing k -space consistency. The first stage learns a set of dictionaries that not only are adaptive to the contrasts, but also capture correlations among multiple contrasts in a sparse transform domain. By capitalizing on the learned dictionaries, the second stage performs coupled sparse coding to remove the aliasing and noise in the corrupted contrasts. The third stage enforces consistency between the denoised contrasts and the measurements in the k -space domain. Numerical experiments, consisting of retrospective under-sampling of various MRI contrasts with a variety of sampling schemes, demonstrate that CDLMRI is capable of capturing structural dependencies between different contrasts. The learned priors indicate notable advantages in multi-contrast MR imaging and promising applications in quantitative MR imaging such as MR fingerprinting.

Reference-Driven Compressed Sensing MR Image Reconstruction Using Deep Convolutional Neural Networks without Pre-Training

Deep learning has proven itself to be able to reduce the scanning time of Magnetic Resonance Imaging (MRI) and to improve the image reconstruction quality since it was introduced into Compressed Sensing MRI (CS-MRI). However, the requirement of using large, high-quality, and patient-based datasets for network training procedures is always a challenge in clinical applications. In this paper, we propose a novel deep learning based compressed sensing MR image reconstruction method that does not require any pre-training procedure or training dataset, thereby largely reducing clinician dependence on patient-based datasets. The proposed method is based on the Deep Image Prior (DIP) framework and uses a high-resolution reference MR image as the input of the convolutional neural network in order to induce the structural prior in the learning procedure. This reference-driven strategy improves the efficiency and effect of network learning. We then add the k-space data correction step to enforce the consistency of the k-space data with the measurements, which further improve the image reconstruction accuracy. Experiments on in vivo MR datasets showed that the proposed method can achieve more accurate reconstruction results from undersampled k-space data.

Sliding motion compensated low-rank plus sparse (SMC-LS) reconstruction for high spatiotemporal free-breathing liver 4D DCE-MRI

Liver dynamic contrast-enhanced MRI (DCE-MRI) requires high spatiotemporal resolution and large field of view to clearly visualize all relevant enhancement phases and detect early-stage liver lesions. The low-rank plus sparse (L + S) reconstruction outperforms standard sparsity-only-based reconstruction through separation of low-rank background component (L) and sparse dynamic components (S). However, the L + S decomposition is sensitive to respiratory motion so that image quality is compromised when breathing occurs during long time data acquisition. To enable high quality reconstruction for free-breathing liver 4D DCE-MRI, this paper presents a novel method called SMC-LS, which incorporates Sliding Motion Compensation into the standard L + S reconstruction. The global superior-inferior displacement of the internal abdominal organs is inferred directly from the undersampled raw data and then used to correct the breathing induced sliding motion which is the dominant component of respiratory motion. With sliding motion compensation, the reconstructed temporal frames are roughly registered before applying the standard L + S decomposition. The proposed method has been validated using free-breathing liver 4D MRI phantom data, free-breathing liver 4D DCE-MRI phantom data, and in vivo free breathing liver 4D MRI dataset. Results demonstrated that SMC-LS reconstruction can effectively reduce motion blurring artefacts and preserve both spatial structures and temporal variations at a sub-second temporal frame rate for free-breathing whole-liver 4D DCE-MRI.

Incorporating reference guided priors into calibrationless parallel imaging reconstruction

PURPOSE

To propose and evaluate a new calibrationless parallel imaging method aimed at further improving the reconstruction accuracy of the accelerated multi-channel MR images.

METHOD

We introduce a new calibrationless parallel imaging method. On top of exploiting joint sparsity cross channels of the target image to be reconstructed, it incorporates similar priors on the grey-level intensity and edge orientation, which both come from a high-spatial resolution reference image that can be easily obtained in many clinical MRI scenarios. The mixed l₂-l₁ norm is used to enforce joint sparsity and a multi-scale gradient operator is applied to extract fine edges from the reference image. Additionally, this optimization problem can be solved via a non-linear conjugate gradient algorithm with line search in this work.

RESULTS

The proposed method is compared with the existing state-of-the-art auto-calibration and calibrationless parallel imaging techniques. The experiments on different in-vivo brain MR datasets show that the proposed method has the superior performance in terms of both artifact suppression and detail preservation.

CONCLUSION

The reference guided calibrationless parallel imaging method can significantly improve the performance of joint reconstruction of target channel images. Even when the reduction factor is high, it can keep edge structures well.

Self-prior image-guided MRI reconstruction with dictionary learning

PURPOSE

A novel method, named self-prior image-guided MRI reconstruction with dictionary learning (SPIDLE), is developed to improve the performance of MR imaging with high acceleration rates. "self-prior" means that the prior image is obtained from the target image itself and any extra MRI scans are not needed.

METHODS

The proposed method integrates self-prior image constraint with compressed sensing (CS) and the dictionary learning (DL) technique. The self-prior image is a preliminary result reconstructed using the undersampled k-space measurements of the target image. Therefore, the self-prior image has similar structural features with the target image, and they match each other accurately. CS approach is applied to the residual error of the target image with the self-prior image, because the error image is much sparser than the target image. The split Bregman method is used to solve the proposed approach to promote fast convergence. For multicoil measurements, each coil image is reconstructed individually and the final result is produced as the square root of sum of squares (SOS) of all channel images.

RESULTS

The performance of the proposed SPIDLE method was inspected using different undersampling schemes and acceleration rates with various types of in vivo MR datasets. Experiments showed that the SPIDLE method is superior to other typical state-of-the-art methods. Specifically, the SPIDLE method produces fewer reconstruction errors, and it is robust to initialization.

CONCLUSIONS

The proposed SPIDLE method substantially widens the applications of prior image-guided MRI reconstruction, especially for applications that are not suitable to use existing MR scans as prior images. The SPIDLE method obviously improves the reconstruction quality for highly undersampled MRI. It is also promising for reconstruction of dynamic MRI and other imaging modalities, such as CT and CT-MRI multimodality imaging.

Fast reference based MRI

In many clinical MRI scenarios, existing imaging information can be used to significantly shorten acquisition time or improve Signal to Noise Ratio (SNR). In those cases, a previously acquired image can serve as a reference image, that may exhibit similarity in some sense to the image being acquired. Examples include similarity between adjacent slices in high resolution MRI, similarity between various contrasts in the same scans and similarity between different scans of the same patients. In this paper we present a general framework for utilizing reference images for fast MRI. We take into account that the reference image may exhibit low similarity with the acquired image and develop a hybrid adaptive-weighted approach for sampling and reconstruction. Experiments demonstrate the performance of the method in three different clinical MRI scenarios: SNR improvement in high resolution brain MRI, utilizing similarity between T2-weighted and fluid-attenuated inversion recovery (FLAIR) for fast FLAIR scanning and utilizing similarity between baseline and follow-up scans for fast follow-up scanning.

Maximum Likelihood Reconstruction for Magnetic Resonance Fingerprinting

This paper introduces a statistical estimation framework for magnetic resonance (MR) fingerprinting, a recently proposed quantitative imaging paradigm. Within this framework, we present a maximum likelihood (ML) formalism to estimate multiple MR tissue parameter maps directly from highly undersampled, noisy k-space data. A novel algorithm, based on variable splitting, the alternating direction method of multipliers, and the variable projection method, is developed to solve the resulting optimization problem. Representative results from both simulations and in vivo experiments demonstrate that the proposed approach yields significantly improved accuracy in parameter estimation, compared to the conventional MR fingerprinting reconstruction. Moreover, the proposed framework provides new theoretical insights into the conventional approach. We show analytically that the conventional approach is an approximation to the ML reconstruction; more precisely, it is exactly equivalent to the first iteration of the proposed algorithm for the ML reconstruction, provided that a gridding reconstruction is used as an initialization.

Accelerated cardiac cine MRI using locally low rank and finite difference constraints

PURPOSE

To evaluate the potential value of combining multiple constraints for highly accelerated cardiac cine MRI.

METHODS

A locally low rank (LLR) constraint and a temporal finite difference (FD) constraint were combined to reconstruct cardiac cine data from highly undersampled measurements. Retrospectively undersampled 2D Cartesian reconstructions were quantitatively evaluated against fully-sampled data using normalized root mean square error, structural similarity index (SSIM) and high frequency error norm (HFEN). This method was also applied to 2D golden-angle radial real-time imaging to facilitate single breath-hold whole-heart cine (12 short-axis slices, 9-13s single breath hold). Reconstruction was compared against state-of-the-art constrained reconstruction methods: LLR, FD, and k-t SLR.

RESULTS

At 10 to 60 spokes/frame, LLR+FD better preserved fine structures and depicted myocardial motion with reduced spatio-temporal blurring in comparison to existing methods. LLR yielded higher SSIM ranking than FD; FD had higher HFEN ranking than LLR. LLR+FD combined the complimentary advantages of the two, and ranked the highest in all metrics for all retrospective undersampled cases. Single breath-hold multi-slice cardiac cine with prospective undersampling was enabled with in-plane spatio-temporal resolutions of 2×2mm(2) and 40ms.

CONCLUSION

Highly accelerated cardiac cine is enabled by the combination of 2D undersampling and the synergistic use of LLR and FD constraints.

MR image reconstruction with block sparsity and iterative support detection

This work aims to develop a novel magnetic resonance (MR) image reconstruction approach motivated by the recently proposed sampling framework with union-of-subspaces model (SUoS). Based on SUoS, we propose a mathematical formalism that effectively integrates a block sparsity constraint and support information which is estimated in an iterative fashion. The resulting optimization problem consists of a data fidelity term and a support detection based block sparsity (SDBS) promoting term penalizing entries within the complement of the estimated support. We provide optional strategies for block assignment, and we also derive unique and robust recovery conditions in terms of the structured restricted isometric property (RIP), namely the block-RIP. The block-RIP constant we derive is lower than that of the previous structured sparse method, which leads to a reduction of the measurements. Simulation results for reconstructing individual and multiple T1/T2-weighted images demonstrate the consistency with our theoretical claims, and show considerable improvement in comparison with methods using only block sparsity or support information.

Compressed Sensing MR Image Reconstruction Exploiting TGV and Wavelet Sparsity

Compressed sensing (CS) based methods make it possible to reconstruct magnetic resonance (MR) images from undersampled measurements, which is known as CS-MRI. The reference-driven CS-MRI reconstruction schemes can further decrease the sampling ratio by exploiting the sparsity of the difference image between the target and the reference MR images in pixel domain. Unfortunately existing methods do not work well given that contrast changes are incorrectly estimated or motion compensation is inaccurate. In this paper, we propose to reconstruct MR images by utilizing the sparsity of the difference image between the target and the motion-compensated reference images in wavelet transform and gradient domains. The idea is attractive because it requires neither the estimation of the contrast changes nor multiple times motion compensations. In addition, we apply total generalized variation (TGV) regularization to eliminate the staircasing artifacts caused by conventional total variation (TV). Fast composite splitting algorithm (FCSA) is used to solve the proposed reconstruction problem in order to improve computational efficiency. Experimental results demonstrate that the proposed method can not only reduce the computational cost but also decrease sampling ratio or improve the reconstruction quality alternatively.