Advertisement

An Edge Enhanced SRGAN for MRI Super Resolution in Slice-Selection Direction

  • Jia Liu
  • Fang Chen
  • Xianyu Wang
  • Hongen LiaoEmail author
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11846)

Abstract

The low resolution MRI in slice-select direction will lead to information loss and artifacts in 2D multi-slices MRI, which is not conducive to the diagnosis and treatment of diseases. Therefore, we proposed an edge enhanced super-resolution generative adversarial networks (EE-SRGAN) for MRI super resolution in slice-select direction. Firstly, a two-stage super-resolution generator network (TSSR) for solving the problem that the down-sampling ratio of MRI resolution in single direction reached 12 times. In addition, in order to overcome the problem of image smoothness caused by high peak signal-to-noise ratio (PSNR) and improve the visual reality of reconstruction image, we construct a generative adversarial networks based on TSSR. Finally, in order to achieve more texture details, we proposed an edge enhanced loss function to optimize the generator network. From the experimental results, we find that our TSSR is better (increased 1.78 dB PSNR), EE-SRGAN provides more satisfactory visual effect and beneficial to segmentation task (increased 2.14% Dice index) than state-of-art super-resolution network.

Keywords

MRI Slice-Selection Two-Stage Super-Resolution Edge enhanced 

References

  1. 1.
    Greenspan, H., Oz, G., Kiryati, N., et al.: MRI inter-slice reconstruction using super-resolution. Magn. Reson. Imaging 20(5), 437–446 (2002)CrossRefGoogle Scholar
  2. 2.
    Duchon, C.E.: Lanczos filtering in one and two dimensions. J. Appl. Meteorol. 18(8), 1016–1022 (1979)CrossRefGoogle Scholar
  3. 3.
    Yang, J., Wright, J., Huang, T.S., et al.: Image super-resolution via sparse representation. IEEE Trans. Image Process. 19(11), 2861–2873 (2010)MathSciNetCrossRefGoogle Scholar
  4. 4.
    Lai, W.S., Huang, J.B., Ahuja, N., et al.: Deep laplacian pyramid networks for fast and accurate super-resolution. In: Proceedings - 30th IEEE Conference on Computer Vision and Pattern Recognition, pp. 5835–5843. IEEE, USA (2017)Google Scholar
  5. 5.
    Kim, J., Kwon Lee, J., Mu Lee, K.: Accurate image super-resolution using very deep convolutional networks. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 1646–1654. IEEE, USA (2016)Google Scholar
  6. 6.
    Pohlen, T., Hermans, A., Mathias, M., Leibe, B.: Full-resolution residual networks for semantic segmentation in street scenes. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4151–4160. IEEE, USA (2017)Google Scholar
  7. 7.
    Ledig, C., et al.: Photo-realistic single image super-resolution using a generative adversarial network. In: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp. 4681–4690. IEEE, USA (2017)Google Scholar
  8. 8.
    Wang, X., Yu, K., Wu, S., Gu, J., Liu, Y., Dong, C., Qiao, Y., Loy, C.C.: ESRGAN: Enhanced Super-Resolution Generative Adversarial Networks. In: Leal-Taixé, L., Roth, S. (eds.) ECCV 2018. LNCS, vol. 11133, pp. 63–79. Springer, Cham (2019).  https://doi.org/10.1007/978-3-030-11021-5_5CrossRefGoogle Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  1. 1.Department of Biomedical Engineering, School of MedicineTsinghua UniversityBeijingChina
  2. 2.Department of Computer Science and EngineeringNanjing University of Aeronautics and AstronauticsNanjingChina

Personalised recommendations