Deriving Neural Network Architectures Using Precision Learning: Parallel-to-Fan Beam Conversion

  • Christopher SybenEmail author
  • Bernhard Stimpel
  • Jonathan Lommen
  • Tobias Würfl
  • Arnd Dörfler
  • Andreas Maier
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11269)


In this paper, we derive a neural network architecture based on an analytical formulation of the parallel-to-fan beam conversion problem following the concept of precision learning. The network allows to learn the unknown operators in this conversion in a data-driven manner avoiding interpolation and potential loss of resolution. Integration of known operators results in a small number of trainable parameters that can be estimated from synthetic data only. The concept is evaluated in the context of Hybrid MRI/X-ray imaging where transformation of the parallel-beam MRI projections to fan-beam X-ray projections is required. The proposed method is compared to a traditional rebinning method. The results demonstrate that the proposed method is superior to ray-by-ray interpolation and is able to deliver sharper images using the same amount of parallel-beam input projections which is crucial for interventional applications. We believe that this approach forms a basis for further work uniting deep learning, signal processing, physics, and traditional pattern recognition.


Machine learning Precision learning Hybrid MRI/X-ray imaging 



This work has been supported by the project P3-Stroke, an EIT Health innovation project. EIT Health is supported by EIT, a body of the European Union.


  1. 1.
    Abadi, M., et al.: TensorFlow: a system for large-scale machine learning. In: OSDI, vol. 16, pp. 265–283 (2016)Google Scholar
  2. 2.
    Christlein, V., et al.: Tutorial: deep learning advancing the state-of-the-art in medical image analysis. In: Maier-Hein, K.H., et al. (eds.) Bildverarbeitung für die Medizin 2017. I, pp. 6–7. Springer, Heidelberg (2017). Scholar
  3. 3.
    Fu, W., et al.: Frangi-Net: a neural network approach to vessel segmentation. In: Maier, A., Deserno, T., Handels, H., Maier-Hein, K., Palm, C., Tolxdorff, T. (eds.) Bildverarbeitung für die Medizin 2018, pp. 341–346 (2018)Google Scholar
  4. 4.
    Huang, Y., Würfl, T., Breininger, K., Liu, L., Lauritsch, G., Maier, A.: Some investigations on robustness of deep learning in limited angle tomography. In: Frangi, A.F., Schnabel, J.A., Davatzikos, C., Alberola-López, C., Fichtinger, G. (eds.) MICCAI 2018. LNCS, vol. 11070, pp. 145–153. Springer, Cham (2018). Scholar
  5. 5.
    Lommen, J., et al.: MR-projection imaging for interventional X/MR-hybrid applications. In: Proceedings of the 49th Annual Meeting of the German Society for Medical Physics (2018)Google Scholar
  6. 6.
    Maier, A.K., et al.: Precision learning: towards use of known operators in neural networks. CoRR abs/1712.00374 (2017).
  7. 7.
    Ramachandran, G., Lakshminarayanan, A.: Three-dimensional reconstruction from radiographs and electron micrographs: application of convolutions instead of fourier transforms. Proc. Nat. Acad. Sci. 68(9), 2236–2240 (1971)MathSciNetCrossRefGoogle Scholar
  8. 8.
    Fahrig, R., et al.: A truly hybrid interventional MR/X-ray system: feasibility demonstration. J. Magn. Reson. Imaging 13(2), 294–300 (2001).\(<\)294::AID-JMRI1042\(>\)3.0.CO;2-XCrossRefGoogle Scholar
  9. 9.
    Shepp, L.A., Logan, B.F.: The fourier reconstruction of a head section. IEEE Trans. Nucl. Sci. 21(3), 21–43 (1974)CrossRefGoogle Scholar
  10. 10.
    Syben, C., et al.: Precision learning: reconstruction filter kernel discretization. In: Noo, F. (ed.) Proceedings of the Fifth International Conference on Image Formation in X-Ray Computed Tomography, pp. 386–390 (2018)Google Scholar
  11. 11.
    Syben, C., Stimpel, B., Leghissa, M., Dörfler, A., Maier, A.: Fan-beam projection image acquisition using MRI. In: Skalej, M., Hoeschen, C. (eds.) 3rd Conference on Image-Guided Interventions & Fokus Neuroradiologie, pp. 14–15 (2017)Google Scholar
  12. 12.
    Wachowicz, K., Murray, B., Fallone, B.: On the direct acquisition of beam’s-eye-view images in MRI for integration with external beam radiotherapy. Phys. Med. Biol. 63(12), 125002 (2018)CrossRefGoogle Scholar
  13. 13.
    Wang, G., et al.: Vision 20/20: simultaneous CT-MRI - next chapter of multimodality imaging. Med. Phys. 42, 5879–5889 (2015). Scholar
  14. 14.
    Würfl, T., Ghesu, F.C., Christlein, V., Maier, A.: Deep learning computed tomography. In: Ourselin, S., Joskowicz, L., Sabuncu, M.R., Unal, G., Wells, W. (eds.) MICCAI 2016. LNCS, vol. 9902, pp. 432–440. Springer, Cham (2016). Scholar
  15. 15.
    Würfl, T., et al.: Deep learning computed tomography: learning projection-domain weights from image domain in limited angle problems. IEEE Trans. Med. Imaging 37(6), 1454–1463 (2018). Scholar
  16. 16.
    Zhu, B., Liu, J.Z., Cauley, S.F., Rosen, B.R., Rosen, M.S.: Image reconstruction by domain-transform manifold learning. Nature 555(7697), 487 (2018)CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Christopher Syben
    • 1
    • 2
    Email author
  • Bernhard Stimpel
    • 1
    • 2
  • Jonathan Lommen
    • 1
    • 2
  • Tobias Würfl
    • 1
  • Arnd Dörfler
    • 2
  • Andreas Maier
    • 1
  1. 1.Pattern Recognition Lab, Department of Computer ScienceFriedrich-Alexander-Universität Erlangen-NürnbergErlangenGermany
  2. 2.Department of NeuroradiologyUniversitätsklinikum Erlangen, Friedrich-Alexander-Universität Erlangen-NürnbergErlangenGermany

Personalised recommendations