Abstract
Using undersampled k-space data for reconstruction is an effective way to accelerate data acquisition of magnetic resonance imaging (MRI). With the development of compressed sensing (CS) theory, many solutions have been proposed for undersampled data reconstruction. Moreover, dictionary learning method has shown good results in improving reconstruction quality. However, CS reconstruction algorithms are time consuming, especially at dictionary training and sparse coding step. The computation overhead is even higher for three-dimensional reconstruction. With a large number of slices, data size can be massive and more time consuming. In this paper, we use three-dimensional dual-dictionary learning scheme for the reconstruction procedure. Three-dimensional dictionaries train the dictionary atoms in image blocks and utilize spatial correlation among MR slices. Dual-dictionary learning method cooperates low-resolution dictionary and high-resolution dictionary for sparse coding and image updating, respectively. Compute unified device architecture (CUDA) is utilized to design the parallel algorithms on graphics processing unit (GPU). We mainly optimize dictionary learning algorithm and image updating. We also develop parallel CPU codes using OpenMP (Open Multi-Processing) and another version of CUDA codes with algorithmic optimization. Experimental results show that more than 324 times of speedup is achieved compared with CPU-only codes with 24 MRI slices and more than 40 times of acceleration compared with OpenMP codes.
Similar content being viewed by others
References
Aharon M, Elad M, Bruckstein A (2006) K-svd: an algorithm for designing overcomplete dictionaries for sparse representation. Signal Process IEEE Trans 54(11):4311–4322
Arabnia HR (1996) Distributed stereo-correlation algorithm. Comput Commun 19(8):707–711
Arabnia HR, Oliver MA (1987) A transputer network for the arbitrary rotation of digitised images. Comput J 30(5):425–432
Arabnia HR, Oliver MA (1989) A transputer network for fast operations on digitised images. In: Computer graphics forum, vol. 8. Blackwell Publishing Ltd, pp 3–11
Davis G, Mallat S, Avellaneda M (1997) Adaptive greedy approximations. Constr Approx 13(1):57–98
Donoho DL (2006) Compressed sensing. Inf Theory IEEE Trans 52(4):1289–1306
Van de Gronde J, Vuçini E (2008) Compressed sensing overview. http://www.cg.tuwien.ac.at/research/publications/2008/Gronde_2008/Gronde_2008-report.pdf
Kreutz-Delgado K, Murray JF, Rao BD, Engan K, Lee TW, Sejnowski TJ (2003) Dictionary learning algorithms for sparse representation. Neural Comput 15(2):349–396
Lustig M, Donoho D, Pauly JM (2007) Sparse mri: The application of compressed sensing for rapid mr imaging. Magn Resonance Med 58(6):1182–1195
Murphy M, Keutzer K, Vasanawala S, Lustig M (2010) Clinically feasible reconstruction time for l1-spirit parallel imaging and compressed sensing mri. In: Proceedings of the ISMRM Scientific Meeting & Exhibition, p 4854
Nvidia: Cublas library (2012). http://docs.nvidia.com/cuda/pdf/CUBLAS_Library.pdf
Nvidia: Cuda c programming guide (2012). http://docs.nvidia.com/cuda/pdf/CUDA_C_Programming_Guide.pdf
Photonics, E.: Cula programmer’s guide release r16a (2012). http://www.culatools.com/files/docs/R16a/CULAReferenceManual_R16a.pdf
Qiu C, Lu W, Vaswani N (2009) Real-time dynamic mr image reconstruction using kalman filtered compressed sensing. In: Acoustics, Speech and Signal Processing, 2009. ICASSP 2009. IEEE International Conference on, IEEE pp 393–396
Rubinstein R, Zibulevsky M, Elad M (2008) Efficient implementation of the k-svd algorithm using batch orthogonal matching pursuit. CS Technion p 40
Shahin M, Tollner E, Evans M, Arabnia H (1999) Watercore features for sorting red delicious apples: a statistical approach. Trans ASAE 42(6):1889–1896
Smith DS, Gore JC, Yankeelov TE, Welch EB (2012) Real-time compressive sensing MRI reconstruction using GPU computing and split Bregman methods. Int J Biomed Imaging 2012:1–6
Song Y, Zhu Z, Lu Y, Liu Q, Zhao J (2014) Reconstruction of magnetic resonance imaging by three-dimensional dual-dictionary learning. Magn Resonance Med 71(2):1285–1298
Sorensen TS, Prieto C, Atkinson D, Hansen MS, Schaeffter T (2010) Gpu accelerated iterative sense reconstruction of radial phase encoded whole-heart mri. In: Proceedings of the 18th scientific meeting, international society for magnetic resonance in medicine. Stockholm, vol. 2869
Stone SS, Haldar JP, Tsao SC, Hwu WM, Sutton BP, Liang ZP et al (2008) Accelerating advanced mri reconstructions on gpus. J Parallel Distrib Comput 68(10):1307–1318
Tosic I, Frossard P (2011) Dictionary learning. Signal Process Mag IEEE 28(2):27–38
Tropp JA (2004) Greed is good: algorithmic results for sparse approximation. Inf Theory IEEE Trans 50(10):2231–2242
Tropp JA, Gilbert AC (2007) Signal recovery from random measurements via orthogonal matching pursuit. Inf Theory IEEE Trans 53(12):4655–4666
Trzasko J, Manduca A (2009) Highly undersampled magnetic resonance image reconstruction via homotopic-minimization. Med Imaging IEEE Trans 28(1):106–121
Acknowledgments
This work is supported by National Natural Science Foundation of China (No. 813716234), National Basic Research Program of China (2010CB834302), and Shanghai Jiao Tong University Medical Engineering Cross Research Funds (YG2013MS30 and YG2011MS51).
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Li, J., Sun, J., Song, Y. et al. Accelerating MRI reconstruction via three-dimensional dual-dictionary learning using CUDA. J Supercomput 71, 2381–2396 (2015). https://doi.org/10.1007/s11227-015-1386-z
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11227-015-1386-z