Abstract
Dynamic magnetic resonance imaging (dMRI) strikes a balance between reconstruction speed and image accuracy in medical imaging field. In this paper, an improved robust tensor principal component analysis (RTPCA) method is proposed to reconstruct the dynamic magnetic resonance imaging (MRI) from highly under-sampled K-space data. The MR reconstruction problem is formulated as a high-order low-rank tenor plus sparse tensor recovery problem, which is solved by robust tensor principal component analysis (RTPCA) with a new tensor nuclear norm (TNN). To further exploit the low-rank structures in multi-way data, the core matrix nuclear norm, extracted from the diagonal elements of the core tensor under tensor singular value decomposition (t-SVD) framework, is also integrated into TNN for enforcing the low-rank structure in MRI datasets. The experimental results show that the proposed method outperforms state-of-the-art methods in terms of both MR image reconstruction accuracy and computational efficiency on 3D and 4D experiment datasets, especially for 4D MR image reconstruction.
Similar content being viewed by others
References
Oost K, Kanneworff AG (2000) Compressed sensing MRI. Int Res Geogr Environ Educ 9:176–180
Verma S, Turkbey B, Muradyan N, Rajesh A, Cornud F, Haider MA, Choyke PL, Harisinghani M (2012) Overview of dynamic contrast-enhanced MRI in prostate cancer diagnosis and management. Am J Roentgenol 198(6):1277–1288
Feng L, Grimm R, Block KT et al (2014) Golden-angle radial sparse parallel MRI: combination of compressed sensing, parallel imaging, and golden-angle radial sampling for fast and flexible dynamic volumetric MRI. Magn Reson Med 72:707–717. https://doi.org/10.1002/mrm.24980
Luo J, Mou Z, Qin B et al (2018) A singular K-space model for fast reconstruction of magnetic resonance images from undersampled data. Med Biol Eng Comput 56:1211–1225. https://doi.org/10.1007/s11517-017-1763-2
Gamper U, Boesiger P, Kozerke S (2008) Compressed sensing in dynamic MRI. Magn Reson Med 59:365–373. https://doi.org/10.1002/mrm.21477
Otazo R, Kim D, Axel L, Sodickson DK (2010) Combination of compressed sensing and parallel imaging for highly accelerated first-pass cardiac perfusion MRI. Magn Reson Med 64:767–776. https://doi.org/10.1002/mrm.22463
Sandilya M, Nirmala SR (2017) Compressed sensing trends in magnetic resonance imaging. Eng Sci Technol an Int J 20:1342–1352. https://doi.org/10.1016/j.jestch.2017.07.001
Lustig M, Donoho D, Pauly JM (2007) Sparse MRI: the application of compressed sensing for rapid MR imaging. Magn Reson Med 58:1182–1195. https://doi.org/10.1002/mrm.21391
Zhong W, Li D, Wang L, Zhang M (2016) Low-rank plus sparse reconstruction using dictionary learning for 3D-MRI. In: 9th international congress on image and signal processing, biomedical engineering and informatics (CISP-BMEI) pp:1407–1411
Xie Q, Zhao Q, Meng D, Xu Z (2017) Kronecker-basis-representation based tensor sparsity and its applications to tensor recovery. IEEE trans on pattern analysis and machine intelligence 40:1888–1902. https://doi.org/10.1109/TPAMI.2017.2734888
Ula SC, Pedro AG et al (2016) Spatio-temporal MRI reconstruction by enforcing local and global regularity via dynamic total and nuclear norm minimization. In: IEEE 13th international symposium on biomedical imaging(ISBI), pp 306–309
Liu S, Cao J, Liu H et al (2018) Group sparsity with orthogonal dictionary and nonconvex regularization for exact MRI reconstruction. Inf Sci 451–452:161–179. https://doi.org/10.1016/j.ins.2018.03.064
Christodoulou AG, Zhang H, Zhao B, Hitchens TK, Ho C, Liang ZP (2013) High-resolution cardiovascular MRI by integrating parallel imaging with low-rank and sparse modeling. IEEE Trans Biomed Eng 60(11):3083–3092
Xu F, Han J, Wang Y et al (2017) Dynamic magnetic resonance imaging via nonconvex low-rank matrix approximation. IEEE Access 5:1958–1966. https://doi.org/10.1109/ACCESS.2017.2657645
Lingala SG, Dibella E, Adluru G (2013) Accelerating free breathing myocardial perfusion MRI using multi-coil radial k − t SLR. Phys Med Biol 58:7309–7327. https://doi.org/10.1088/0031-9155/58/20/7309
Trémoulhéac B, Dikaios N, Atkinson D et al (2014) Dynamic MR image reconstruction-separation from undersampled (k, t)-space via low-rank plus sparse prior. IEEE transactions on medical imaging 33(8):1689–1701
Otazo R, Sodickson DK, Cande E (2015) Low-rank plus sparse matrix decomposition for accelerated dynamic MRI with separation of background and dynamic components. Magn Reson Med 73:1125–1136. https://doi.org/10.1002/mrm.25240
Zhao B, Haldar JP, Christodoulou AG et al (2012) Image reconstruction from highly undersampled (k, t)-space data with joint partial separability and sparsity constraints. IEEE transactions on medical imaging 31(9):1809–1820
Lingala SG, Jacob M (2013) Blind compressive sensing dynamic MRI. IEEE Trans Med Imaging 32:1132–1145
Rahmani M, Atia GK (2017) High dimensional low rank plus sparse matrix decomposition. IEEE Trans Signal Process 65:2004–2019
Kolda TG, Bader BW (2009) Tensor decompositions and applications. SIAM Rev 51:455–500
Rauhut H, Schneider R, Stojanac Ž (2017) Low rank tensor recovery via iterative hard thresholding. Linear Algebra Appl 523:220–262. https://doi.org/10.1016/j.laa.2017.02.028
Sidiropoulos ND, Lathauwer D et al (2017) Tensor decomposition for signal processing and machine learning. IEEE Trans Signal Process 65:3551–3582. https://doi.org/10.1145/242224.242229
Ying J, Lu H, Wei Q et al (2017) Hankel matrix nuclear norm regularized tensor completion for N-dimensional exponential signals. IEEE Trans Signal Process 65(14):3702–3717
An J, Zhang X, Zhou H, Jiao L (2018) Tensor-based low-rank graph with multimanifold regularization for dimensionality reduction of hyperspectral images. IEEE Trans Geosci Remote Sens 56:4731–4746. https://doi.org/10.1109/TGRS.2018.2835514
Derksen H (2016) On the nuclear norm and the singular value decomposition of tensor. Found Comput Math 16:779–811. https://doi.org/10.1007/s10208-015-9264-x
Liu Y, Member S, Long Z, Zhu C (2018) Image completion using low tensor tree rank and total variation minimization. IEEE Trans Multimed:1. https://doi.org/10.1109/TMM.2018.2859026
Xu Y, Wu Z et al (2018) Joint reconstruction and anomaly detection from compressive hyperspectral images using Mahalanobis distance-regularized tensor RPCA. IEEE Trans Geosci Remote Sens 56:2919–2930
Zhou P, Lu C, Member S et al (2018) Tensor factorization for low-rank tensor completion. IEEE Trans Image Process 27:1152–1163
He J, Liu Q, Christodoulou AG, Ma C, Lam F, Liang ZP (2016) Accelerated high-dimensional MR imaging with sparse sampling using low-rank tensors. IEEE Trans Med Imaging 35(9):2119–2129
Huang J, Zhou G, Yu G (2019) Orthogonal tensor dictionary learning for accelerated dynamic MRI. Med Biol Eng Comput 57:1933–1946. https://doi.org/10.1007/s11517-019-02005-x13
Yu Y, Jin J, Liu F, Crozier S (2014) Multidimensional compressed sensing MRI using tensor decomposition-based sparsifying transform. PLoS One 9:1–13. https://doi.org/10.1371/journal
Roohi SF, Zonoobi D, Kassim AA, Jaremko JL (2017) Multi-dimensional low rank plus sparse decomposition for reconstruction of under-sampled dynamic MRI. Pattern Recogn 63:667–679. https://doi.org/10.1016/j.patcog.30.31.32
Lu C, Feng J, Chen Y et al (2016) Tensor robust principal component analysis: exact recovery of corrupted low-rank tensors via convex optimization. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit:5249–5257. https://doi.org/10.1109/CVPR.2016.567
Chen L, Liu Y, Zhu C (2018) Robust tensor principal component analysis in all modes. In: 2018 IEEE Int Conf Multimed Expo, pp 1–6
Liu Y, Member S, Chen L, Zhu C (2018) Improved robust tensor principle component analysis via low rank core matrix. IEEE J Sel Top Signal Process:1. https://doi.org/10.1109/JSTSP.2018.2873142
Zhang C, Hu W, Jin T, Mei Z (2018) Nonlocal image denoising via adaptive tensor nuclear norm minimization. Neural Comput Applic 29:3–19. https://doi.org/10.1007/s00521-015-2050-5
Lu C, Feng J, Chen Y et al (2019) Tensor robust principle component analysis with a new tensor nuclear norm. IEEE Trans Pattern Anal Mach Intell
Cao W, Wang Y, Sun J, Meng D, Yang C, Cichocki A, Xu Z (2016) Total variation regularized tensor RPCA for background subtraction from compressive measurements. IEEE Trans Image Process 25:4075–4090. https://doi.org/10.1109/TIP.2016.2579262
Rasch J, Kolehmainen V, Nivajarvi R et al (2018) Dynamic MRI reconstruction from undersampled data with an anatomical prescan. Inverse Probl 34:074001. https://doi.org/10.1088/1361-6420/aac3af
Wang S, Ke Z, Cheng H et al (2018) DIMENSION: dynamic MR imaging with both k-space and spatial prior knowledge obtained via multi-supervised network training. NMR Biomed. https://doi.org/10.1002/nbm.4131
Wang, S, Chen, Y, Xiao, T, Ke, Z, Liu, Q, & Zheng, H (2019) LANTERN: learn analysis transform network for dynamic magnetic resonance imaging with small dataset. arXiv preprint arXiv:1908.09140
Acknowledgments
This work is supported in part by the National Natural Science Foundation of China (61672466 and 61671405), Joint Fund of Zhejiang Provincial Natural Science Foundation (LSZ19F010001), the Key Research and Development Program of Zhejiang Province (2020C03060), Natural Science Foundation of Zhejiang Province (LY18D060009), and this work is also supported by the 521 Talents project of Zhejiang Sci-Tech University.
Author information
Authors and Affiliations
Corresponding author
Additional information
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendices
Appendix I. Robust principal component analysis model based on matrix and tensor
The RPCA decomposes a matrix \( X\in {\mathrm{R}}^{N_1\times {N}_2} \) into a low-rank matrix L and a sparse component S, and forms the following convex optimization problem:
where ‖L‖∗ is the matrix nuclear norm, which is the sum of the singular values of matrix L, and ‖S‖1 denotes the l1 norm of matrix S.
The RTPCA decomposes 3-order tensor \( \chi \in {\mathrm{R}}^{N_1\times {N}_2\times {N}_3} \) into a low-rank tensor and sparse tensor as follows: χ = ℓ + ξ, where ℓ and ξ denote the low-rank component and sparse component respectively. Thus,
RPCA is classical principal component analysis that has been widely used for data processing problems. However, it is based on matrix mode. It is natural to consider extending RPCA to manipulate the tensor data by taking advantages of its multi-dimensional structure. Robust tensor principal component analysis extracts the low-rank and sparse component of multi-dimensional data by t-SVD, which can be used for many data analysis problems. RTPCA has been applied to three groups of numerical experiments on image denoising, face images, and motion separation in videos.
Appendix II The tensor singular value decomposition operator
Given a 3-order tensor \( \chi \in {\mathrm{R}}^{N_1\times {N}_2\times {N}_3} \), N1, N2, and N3 are the 1th, 2th, and 3thdimensions of χ. The t-SVD is given by χ = U ∗ S ∗ VH, where U and V are orthogonal tensors of size N1 × N1 × N3 and N2 × N2 × N3, respectively, and S is an f-diagonal tensor. This t-SVD is constructed as follows: tensor χ is transformed into Fourier domain, and let \( \overline{\chi}\left(:,:,i\right)={\overline{U}}^{(i)}{\overline{S}}^{(i)}{\left({\overline{V}}^{(i)}\right)}^H \) be matrix singular value decomposition for frontal slices \( i=1,\dots, \left\lceil \frac{N_3+1}{2}\right\rceil \). For frontal slices \( i=\left\lceil \frac{N_3+1}{2}\right\rceil +1,\dots, {N}_3, \)let\( {\overline{U}}^{(i)}=\mathrm{conj}\left({\overline{U}}^{\left({N}_3-i+2\right)}\right) \),\( {\overline{S}}^{(i)}={\overline{S}}^{\left({N}_3-i+2\right)} \), and \( {\overline{V}}^{(i)}==\mathrm{conj}\left({\overline{V}}^{\left({N}_3-i+2\right)}\right). \) This processing scheme on all frontal slices is consistent with tensor-tensor product mode. Notice that the work reduces the number of singular value decomposition (SVD) to \( \left\lceil \frac{N_3+1}{2}\right\rceil \), while the previous works require computing SVD N3times; therefore, computation efficiency of this novel tensor nuclear norm is improved significantly when N3 is large [37].
T-SVD method for the 3-order tensor | |
Input: \( \chi \in {\mathrm{R}}^{N_1\times {N}_2\times {N}_3} \) 1. Compute \( \overline{\chi}\leftarrow \mathrm{fft}\left(\chi, \left[\right],3\right), \) 2. Compute each frontal slice of \( \overline{U},\kern0.5em \overline{S} \) and \( \overline{V} \) from the \( \overline{\chi} \) by for i = 1, ... , to \( \left\lceil \frac{N_3+1}{2}\right\rceil \) do [\( {\overline{\mathrm{U}}}^{\left(\mathrm{i}\right)},{\overline{\mathrm{S}}}^{\left(\mathrm{i}\right)},{\overline{\mathrm{V}}}^{\left(\mathrm{i}\right)} \)] = SVD(χ(:,  : , i)), end for for \( i=\left\lceil \frac{N_3+1}{2}\right\rceil +1,...,{N}_3 \) do \( {\overline{\mathrm{U}}}^{\left(\mathrm{i}\right)}=\mathrm{conj}\left({\overline{\mathrm{U}}}^{\left({N}_3-\mathrm{i}+2\right)}\right); \) \( {\overline{\mathrm{S}}}^{\left(\mathrm{i}\right)}={\overline{\mathrm{S}}}^{\left({N}_3-\mathrm{i}+2\right)}; \) \( {\overline{\mathrm{V}}}^{\left(\mathrm{i}\right)}=\mathrm{conj}\left({\overline{\mathrm{V}}}^{\left({N}_3-\mathrm{i}+2\right)}\right); \) end for 3. \( U\leftarrow \mathrm{ifft}\left(\overline{U},\left[\right],3\right),S\leftarrow \mathrm{ifft}\left(\overline{S},\left[\right],3\right),V\leftarrow \mathrm{ifft}\left(\overline{V},\left[\right],3\right), \) Output: U, S, V |
Rights and permissions
About this article
Cite this article
Jiang, M., Shen, Q., Li, Y. et al. Improved robust tensor principal component analysis for accelerating dynamic MR imaging reconstruction. Med Biol Eng Comput 58, 1483–1498 (2020). https://doi.org/10.1007/s11517-020-02161-5
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11517-020-02161-5