Simultaneous use of individual and joint regularization terms in compressive sensing: Joint reconstruction of multi‐channel multi‐contrast MRI acquisitions

Abstract

Multi‐contrast images are commonly acquired together to maximize complementary diagnostic information, albeit at the expense of longer scan times. A time‐efficient strategy to acquire high‐quality multi‐contrast images is to accelerate individual sequences and then reconstruct undersampled data with joint regularization terms that leverage common information across contrasts. However, these terms can cause features that are unique to a subset of contrasts to leak into the other contrasts. Such leakage‐of‐features may appear as artificial tissues, thereby misleading diagnosis. The goal of this study is to develop a compressive sensing method for multi‐channel multi‐contrast magnetic resonance imaging (MRI) that optimally utilizes shared information while preventing feature leakage. Joint regularization terms group sparsity and colour total variation are used to exploit common features across images while individual sparsity and total variation are also used to prevent leakage of distinct features across contrasts. The multi‐channel multi‐contrast reconstruction problem is solved via a fast algorithm based on Alternating Direction Method of Multipliers. The proposed method is compared against using only individual and only joint regularization terms in reconstruction. Comparisons were performed on single‐channel simulated and multi‐channel in‐vivo datasets in terms of reconstruction quality and neuroradiologist reader scores. The proposed method demonstrates rapid convergence and improved image quality for both simulated and in‐vivo datasets. Furthermore, while reconstructions that solely use joint regularization terms are prone to leakage‐of‐features, the proposed method reliably avoids leakage via simultaneous use of joint and individual terms, thereby holding great promise for clinical use.

Joint calibrationless reconstruction of highly undersampled multicontrast MR datasets using a low‐rank Hankel tensor completion framework

“…Several recent multicontrast reconstruction approaches 22,23,49 have used image structural correlation based on compressed sensing and made improvements over single-contrast reconstruction. In contrast, the proposed MC-HTC approach exploits the low-rank characteristic of multicontrast tensor for reconstruction without requiring calibration or additional prior information.…”

“…To further exploit this information redundancy, parallel imaging has been combined with compressed sensing reconstruction 13,14 for additional regularizations on similar or consistent structural edges across different contrasts. [15][16][17][18][19][20][21][22][23] Additionally, the highly correlated anatomical structure among multicontrast images ensures strong image-content similarity. This image structural correlation has been explored as a locally low-rank constraint, 24 demonstrating applications in parameter mapping, 25 dynamic imaging, 26 and multicontrast image denoising.…”

“…However, by incorporating complementary undersampling patterns, the coherency of artifacts can be mitigated as revealed by some compressed sensing reconstruction approaches. 13,14,23 In this study, we have further enhanced the complementary sampling strategy by orthogonally alternating the phase-encoding directions among contrasts, forming "pseudo-2D" sampling patterns in multicontrast data acquisition. The effects of pseudo-2D undersampling were shown by the obvious leakage of artifacts from other contrasts for the first few iterations in the MC-HTC reconstruction (Supporting Information Figure S5) and the estimated 2D common spatial support (Supporting Information Figure S6).…”

Accelerated high-resolution free-breathing 3D whole-heart T2-prepared black-blood and bright-blood cardiovascular magnetic resonance

“…(1) to exploit the structural similarity (sparsity) across contrasts and jointly reconstruct T 2 Prep-IR and T 2 Prep BOOST images. This approach could potentially provide higher quality images than reconstructing the T 2 Prep-IR and T 2 Prep BOOST images separately [33][34][35].…”

Trends in Compressive Sensing for EEG Signal Processing Applications

“…It was then developed for many other applications such as ultrasound imaging [ 5 ], face recognition [ 6 , 7 ], single- pixel camera [ 8 ], wireless sensors networks [ 9 , 10 ], cognitive radio networks [ 11 , 12 ], sound localization [ 13 ], audio processing [ 14 , 15 ], radar imaging [ 16 , 17 ], image processing [ 18 , 19 ], and video processing [ 20 , 21 ]. Similarly, CS has contributed to various neural engineering research including, neuronal network connectivity [ 22 ], magnetic resonance image (MRI) acquisition [ 23 ], MRI reconstruction [ 24 ], electroencephalogram (EEG) monitoring [ 25 ], compressive imaging [ 26 , 27 ], and other applications.…”