Metal-Artifact Reduction Using Deep-Learning Based Sinogram Completion: Initial Results

Bernhard E. H. Claus, Yannan Jin, Lars A. Gjesteby, Ge Wang, Bruno De Man


Published in:Fully3D 2017 Proceedings


CT, computed tomography, deep learning, metal artifacts, MAR, missing data estimation, Sinogram completion, Sinogram interpolation
Metal artifact reduction (MAR) remains a hard problem with remaining limitations in image quality after more than three decades of study. Recent successes in deep learning achieve state-of-the-art performance for a range of extremely difficult and complex problems. In this work an application of deep learning to the problem of metal artifact reduction via estimation of missing data in the sinogram domain is investigated. Initial results obtained with projection data simulated from simple geometric objects show promising improvements in the quality of the estimated data in the sinogram, which is visually clearly superior to a linear interpolation approach. This improvement also translates into reduced streaking and banding artifacts in the corresponding reconstructed images. These encouraging results clearly demonstrate the potential of dee p learning for addressing MAR and similar missing data problems
Bernhard E. H. Claus
General Electric Global Research Center
Yannan Jin
General Electric Global Research Center
Lars A.Gjesteby
Rensselaer Polytechnic Institute (RPI)
Ge Wang
Rensselaer Polytechnic Institute (RPI)
Bruno De Man
General Electric Global Research Center
  1. Krizhevsky A, Sutskever I, Hinton G. 2012 ImageNet classification with deep convolutional neural networks. In Proc. 26th Annual Conf. on Neural Information Processing Systems, Lake Tahoe, NV, 3–6 December 2012, pp. 1090–1098.
  2. Hinton G et al. 2012 Deep neural networks for acoustic modeling in speech recognition. IEEE Signal Process. Mag. 29, 82–97.
  3. Yann LeCun, Yoshua Bengio& Geoffrey Hinton, Deep learning, Nature 521, 436–444 (28 May 2015).
  4. Wang, G. “A Perspective on Deep Imaging”, IEEE Access, vol 4., November 2016.
  5. Gjesteby, L., De Man, B., Jin, Y., Paganetti, H., Verburg, J., Giantsoudi, D., Wang, G., Metal Artifacrt Reduction in CT: Where Are We After Four Decades?, IEEE Access, vol. 4, 2016.
  6. J. Y. Huang et al., ‘‘An evaluation of three commercially available metal artifact reduction methods for CT imaging,’’ Phys. Med. Biol., vol. 60, no. 3, p. 1047, 2015.
  7. Wagenaar D, van der Graaf ER, van der Schaaf A, Greuter MJW (2015) Quantitative Comparison of Commercial and Non-Commercial Metal Artifact Reduction Techniques in Computed Tomography. PLoS ONE 10(6), 2015.
  8. J. Müller and T. M. Buzug, ``Spurious structures created by interpolationbased CT metal artifact reduction,'' in Proc. SPIE 7258, Med. Imag., Phys. Med. Imag., Mar. 2009.
  9. Kalender W. et al. "Reduction of CT artifacts caused by metallic impants." Radiology, 1987.
  10. Glover G. and Pelc N. "An algorithm for the reduction of metal clip artifacts in CT reconstructions." Med. Phys., 1981.
  11.  Mahnken A. et al, "A new algoritbm for metal artifact reduction in computed tomogrpaby, In vitro and in vivo evaluation after total hip replacement." Investigative Radiology, vol. 38, no. 12, pp. 769-775, Dec. 2003.
  12.  Meyer E, Raupach R, Lell M, Schmidt B, Kachelriess M., Normalized metal artifact reduction (NMAR) in computed tomography. Med Phys. 2010 Oct;37(10):5482-93.
  13. Boas FE, Fleischmann D., Evaluation of two iterative techniques for reducing metal artifacts in computed tomography. Radiology. 2011 Jun;259(3):894-902.
  14. Meyer E, Raupach R, Lell M, Schmidt B, Kachelrieß M. Frequency split metal artifact reduction (FSMAR) in computed tomography. Med Phys. 2012;39: 1904–16.
  15. Dong C., Loy C.C., He, K., Tang, X., Image Super-Resolution Using Deep Convolutional Networks, IEEE Trans. PAMI, 38(2), 295-307, 2016.
  16. Kim, J., Lee, J.K., and Lee, K.M., Accurate Image Super-Resolution Using Very Deep Convolutional Networks, IEEE CVPR2016, 1646- 1654.