Advertisement

Dictionary Learning-Based MR Image Reconstruction in the Presence of Speckle Noise: Greedy Versus Convex

  • M. V. R. ManimalaEmail author
  • C. Dhanunjaya Naidu
  • M. N. Giri Prasad
Conference paper
Part of the Advances in Intelligent Systems and Computing book series (AISC, volume 898)

Abstract

The trend of acquiring samples at a rate far below Nyquist rate is termed as compressive sensing (CS). CS enables the accurate recovery of signals/images by exploiting the underlying sparsity in either transform or signal domain. CS is useful in reducing acquisition time in MR imaging. In this paper, CS based on dictionary learning has been applied for MR imaging in the presence of speckle noise. Performance of two classes of sparse recovery techniques, namely convex optimization and greedy iterative algorithms, has been investigated and compared, when employed for the sparse coding stage of an adaptive patch-based dictionary. Two greedy algorithms, orthogonal matching pursuit (OMP) and compressive sampling matching pursuit (CoSaMP), are considered and contrasted with convex techniques: basis pursuit (BP) and least absolute shrinkage and selection operator (LASSO). Experimentation has been done on spine, knee, and brain image by varying the speckle noise variance, sparsity threshold per patch for various sampling schemes. Results show that convex techniques achieve higher peak signal-to-noise ratio (PSNR) than greedy algorithms in the presence of high noise.

Keywords

Compressive sensing Orthogonal matching pursuit Compressive sampling matching pursuit Basis pursuit Least absolute shrinkage and selection operator 

Notes

Declaration

The brain image and various masks used in this work are made available online by Dr. Saiprasad Ravishankar and Dr. Yoram Bresler for research purpose, and the authors would like to thank them for the same.

References

  1. 1.
    Lustig, M., Donoho, D.L., Santos, J.M., Pauly, J.M.: Compressed sensing MRI. IEEE Signal Process. Mag. 25(2), 72–82 (2008)CrossRefGoogle Scholar
  2. 2.
    Ravishankar, S., Bresler, Y.: MR Image reconstruction from highly undersampled k-space data by dictionary learning. IEEE Trans. Med. Imaging 30(5), 1028–1041 (2011)CrossRefGoogle Scholar
  3. 3.
    Aharon, M., Elad, M., Bruckstein, A.: K-SVD: an algorithm for designing overcomplete dictionaries for sparse representation. IEEE Trans. Signal Process. 54(11), 4311–4322 (2006)CrossRefGoogle Scholar
  4. 4.
    Khaled, Z., et al.: Real-time speckle reduction and coherence enhancement in ultrasound imaging via nonlinear anisotropic diffusion. IEEE Trans. Biomed. Eng. 49(9), 997–1014 (2002)CrossRefGoogle Scholar
  5. 5.
    Tropp, J., Gilbert, A.C.: Signal recovery from random measurements via orthogonal matching pursuit. IEEE Trans. Inf. Theory 53(12), 4655–4666 (2007)MathSciNetCrossRefGoogle Scholar
  6. 6.
    Needell, D., Tropp, J.A.: CoSaMP: iterative signal recovery from incomplete and inaccurate samples. Commun. ACM 53(12), 93–100 (2010)CrossRefGoogle Scholar
  7. 7.
    Cheny, S.S., et al.: Atomic decomposition by basis pursuit. Soc. Ind. Appl. Math. 20(1), 33–61 (1998)MathSciNetCrossRefGoogle Scholar
  8. 8.
    Tibshirani, R.: Regression shrinkage and selection via the lasso. J. R. Statist. Soc. B 58, 267–288 (1996)MathSciNetzbMATHGoogle Scholar
  9. 9.
    Ravishankar, S., Bresler, Y.: Learning sparsifying transforms. IEEE Trans. Signal Process. 61(5), 1072–1086 (2013)MathSciNetCrossRefGoogle Scholar
  10. 10.
    Manimala, M.V.R., et al.: Sparse recovery algorithms based on dictionary learning for MR image reconstruction. In: IEEE 2016 International Conference on Wireless Networks, Signal Processing and Networking, pp. 1354–1360, March (2016)Google Scholar
  11. 11.
    Candès, E., Romberg, J.: \( \mathop l\nolimits_{1} \)-magic: recovery of sparse signals via convex programming. Caltech, October (2005)Google Scholar
  12. 12.
    Donoho, D., et al.: About SparseLab, Stanford University, Version.100, May (2006)Google Scholar
  13. 13.
  14. 14.

Copyright information

© Springer Nature Singapore Pte Ltd. 2019

Authors and Affiliations

  • M. V. R. Manimala
    • 1
    Email author
  • C. Dhanunjaya Naidu
    • 2
  • M. N. Giri Prasad
    • 1
  1. 1.JNTUAAnanthapuramuIndia
  2. 2.VNRVJIETHyderabadIndia

Personalised recommendations