Whole Heart and Great Vessel Segmentation with Context-aware of Generative Adversarial Networks

  • Mina Rezaei
  • Haojin Yang
  • Christoph Meinel
Conference paper
Part of the Informatik aktuell book series (INFORMAT)


Automatic segmentation of cardiac magnetic resonance imaging (CMRI) is an important application in clinical tasks. However, semantic segmentation of the myocardium and blood pool in CMR is a challenge due to differentiating branchy structures and slicing fuzzy boundaries. In this paper, we propose an automatic deep architecture for simultaneous myocardium and blood pool segmentation on patients with congenital heart disease (CHD). Inspired by vanilla generative adversarial networks (GANs), we propose a cascade of conditional GANs for semantic segmentation. The proposed cascade has three stages that are designed to share convolutional features and weights. Each stage has a conditional generative adversarial network with a unique loss function and trains on different images from the same patients. We further apply AutoContext Model to implement a context-aware generative adversarial network. The proposed method evaluated on the HVSMR dataset and the experimental results demonstrated the superior performance of our approach.


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    LeCun Y, Bengio Y, Hinton G. Deep learning. Nature. 2015;521(7553):436–444.Google Scholar
  2. 2.
    Mirza M, Osindero S. Conditional generative adversarial nets. Comp Res Rep. 2014;abs-1411.1784.Google Scholar
  3. 3.
    Reed SE, Akata Z, Mohan S, et al. Learning what and where to draw. In: Lee DD, Sugiyama M, Luxburg UV, et al., editors. Advances in Neural Information Processing Systems 29; 2016. p. 217–225.Google Scholar
  4. 4.
    Wang X, Shrivastava A, Gupta A. A-fast-rcnn: hard positive generation via adversary for object detection. arXiv preprint arXiv:170403414. 2017.
  5. 5.
    Isola P, Zhu J, Zhou T, et al. Image-to-image translation with conditional adversarial networks. Comp Res Rep. 2016;abs/1611.07004.Google Scholar
  6. 6.
    Xue Y, Xu T, Zhang H, et al. SegAN: adversarial network with multi-scale Loss for medical image segmentation. Comp Res Rep. 2017;abs/1706.01805.Google Scholar
  7. 7.
    Ronneberger O, Fischer P, Brox T. U-net: Convolutional networks for biomedical image segmentation. Proc MICCAI. 2015; p. 234–241.Google Scholar
  8. 8.
    Moeskops P, Veta M, Lafarge MW, et al. Adversarial training and dilated convolutions for brain MRI segmentation. Comp Res Rep. 2017;abs/1707.03195.Google Scholar
  9. 9.
    Kohl S, Bonekamp D, Schlemmer H, et al. Adversarial networks for the detection of aggressive prostate cancer. Comp Res Rep. 2017;abs/1702.08014.Google Scholar
  10. 10.
    Zhu W, Xie X. Adversarial deep structural networks for mammographic mass segmentation. Comp Res Rep. 2016;abs/1612.05970.Google Scholar
  11. 11.
    Radford A, Metz L, Chintala S. Unsupervised representation learning with deep convolutional generative adversarial networks. Comp Res Rep. 2015;abs/1511.06434.Google Scholar
  12. 12.
    Shahzad R, Gao S, Tao Q, et al. In: Automated cardiovascular segmentation in patients with congenital heart disease from 3D CMR scans: combining multiatlases and level-sets. Springer International Publishing; 2017. p. 147–155.Google Scholar
  13. 13.
    Yu L, Yang X, Qin J, et al. In: 3D FractalNet: dense volumetric segmentation for cardiovascular MRI volumes; 2017. p. 103–110.Google Scholar
  14. 14.
    Wolterink JM, Leiner T, Viergever MA, et al.; Springer. Dilated convolutional neural networks for cardiovascular MR segmentation in congenital heart disease. International Workshop on Reconstruction and Analysis of Moving Body Organs. 2016; p. 95–102.Google Scholar

Copyright information

© Springer-Verlag GmbH Deutschland 2018

Authors and Affiliations

  1. 1.Hasso-Plattner Institute for Digital EngineeringPotsdamDeutschland

Personalised recommendations