ISLES Challenge: U-Shaped Convolution Neural Network with Dilated Convolution for 3D Stroke Lesion Segmentation

  • Alzbeta TureckovaEmail author
  • Antonio J. Rodríguez-Sánchez
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 11383)


In this paper, we propose the algorithm for stroke lesion segmentation based on a deep convolutional neural network (CNN). The model is based on U-shaped CNN, which has been applied successfully to other medical image segmentation tasks. The network architecture was derived from the model presented in Isensee et al. [1] and is capable of processing whole 3D images. The model incorporates the convolution layers through upsampled filters – also known as dilated convolution. This change enlarges filter’s field of the view and allows the net to integrate larger context into the computation. We add the dilated convolution into different parts of network architecture and study the impact on the overall model performance. The best model which uses the dilated convolution in the input of the net outperforms the original architecture in nearly all used evaluation metrics. The code and trained models can be found on the GitHub website:


Medical image segmentation Deep convolutional neural networks U-Net Dilated convolution 



This work was supported by the Internal Grant Agency of Tomas Bata University under the Project no. IGA/CebiaTech/2018/003 and further by resources of A. I. Lab ( and IIS group at the University of Innsbruck (


  1. 1.
    Isensee, F., Kickingereder, P., Wick, W., Bendszus, M., Maier-Hein, K.H.: Brain tumor segmentation and radiomics survival prediction: contribution to the BRATS 2017 challenge. In: Crimi, A., Bakas, S., Kuijf, H., Menze, B., Reyes, M. (eds.) BrainLes 2017. LNCS, vol. 10670, pp. 287–297. Springer, Cham (2018). Scholar
  2. 2.
    Global Health Estimates 2016: Deaths by cause, sex, by country and by region, 2000–2016. World Health Organization, Geneva (2018)Google Scholar
  3. 3.
    Maier, O., et al.: ISLES 2015: a public evaluation benchmark for ischemic stroke lesion segmentation from multispectral MRI. Med. Image Anal. (2016). ISSN 1361-8415CrossRefGoogle Scholar
  4. 4.
    Kistler, M., Bonaretti, S., Pfahrer, M., Niklaus, R., Büchler, P.: The virtual skeleton database: an open access repository for biomedical research and collaboration. JMIR (2013). Scholar
  5. 5.
    Usinskas, A., Gleizniene, R.: Ischemic stroke region recognition based on ray tracing. In: Proceedings of International Baltic Electronics Conference (2006).
  6. 6.
    Rajini, N.H., Bhavani, R.: Computer aided detection of ischemic stroke using segmentation and texture features. Measurement 46, 1865–1874 (2013)CrossRefGoogle Scholar
  7. 7.
    Maier, O., Schröder, C., Forkert, N.D., Martinetz, T., Handels, H.: Classifiers for ischemic stroke lesion segmentation: a comparison study. PLoS One 10(12), e0145118 (2015). Scholar
  8. 8.
    Winzeck, S., et al.: ISLES 2016 and 2017-benchmarking ischemic stroke lesion outcome prediction based on multispectral MRI. Front. Neurol. 9 (2018). ISSN 1664-2295
  9. 9.
    Chen, L.-Ch., Papandreou, G., Kokkinos, I., Murphy, K., Yuille, A.L.: DeepLab: semantic image segmentation with deep convolutional nets, atrous convolution, and fully connected CRFs. In: CoRR (2016)Google Scholar
  10. 10.
    Yu, F.; Koltun, V.: Multi-scale context aggregation by dilated convolutions. In: CoRR (2015)Google Scholar
  11. 11.
    Ronneberger, O., Fischer, P., Brox, T.: U-Net: convolutional networks for biomedical image segmentation. In: Navab, N., Hornegger, J., Wells, W.M., Frangi, A.F. (eds.) MICCAI 2015. LNCS, vol. 9351, pp. 234–241. Springer, Cham (2015). Scholar
  12. 12.
    Amorim, P.H.A., et al.: 3D U-Nets for brain tumor segmentation in MICCAI 2017 BraTS challenge. In: 2017 International MICCAI BraTS Challenge (2017)Google Scholar
  13. 13.
    Gordienko, Y., et al.: Deep learning with lung segmentation and bone shadow exclusion techniques for chest x-ray analysis of lung cancer. In: CoRR (2017)Google Scholar
  14. 14.
    Novikov, A.A., et al.: Fully convolutional architectures for multi-class segmentation in chest radiographs. In: CoRR (2017)Google Scholar
  15. 15.
    Tustison, N.J., et al.: N4ITK: improved N3 bias correction. IEEE Trans. Med. Imaging 29(6), 1310–1320 (2010). Scholar
  16. 16.
    Ulyanov, D., Vedaldi, A., Lempitsky, V.S.: Instance normalization: the missing ingredient for fast stylization. In: CoRR (2016)Google Scholar
  17. 17.
    Dice, L.R.: Measures of the amount of ecologic association between species. Ecology. 26(3), 297–302 (1945). Scholar
  18. 18.
    Rockafellar, R.T., Wets, R.J.-B.: Variational Analysis, p. 117. Springer, Heidelberg (2005). ISBN 3-540-62772-3CrossRefGoogle Scholar
  19. 19.
    Olson, D.L., Delen, D.: Advanced Data Mining Techniques, 1st edn, p. 138. Springer, Heidelberg (2008). ISBN 3-540-76916-1CrossRefzbMATHGoogle Scholar
  20. 20.
    Taha, A.A., Hanbury, A.: Metrics for evaluating 3D medical image segmentation: analysis, selection, and tool. BMC Med. Imaging 15, 29 (2015). Scholar

Copyright information

© Springer Nature Switzerland AG 2019

Authors and Affiliations

  • Alzbeta Tureckova
    • 1
    Email author
  • Antonio J. Rodríguez-Sánchez
    • 2
  1. 1.Faculty of Applied InformaticsTomas Bata University in ZlinZlinCzech Republic
  2. 2.Intelligent and Interactive Systems, Institute of Computer ScienceUniversity of InnsbruckInnsbruckAustria

Personalised recommendations