Brain Tumor Segmentation Using a Generative Model with an RBM Prior on Tumor Shape

  • Mikael Agn
  • Oula Puonti
  • Per Munck af Rosenschöld
  • Ian Law
  • Koen Van Leemput
Part of the Lecture Notes in Computer Science book series (LNCS, volume 9556)


In this paper, we present a fully automated generative method for brain tumor segmentation in multi-modal magnetic resonance images. The method is based on the type of generative model often used for segmenting healthy brain tissues, where tissues are modeled by Gaussian mixture models combined with a spatial atlas-based tissue prior. We extend this basic model with a tumor prior, which uses convolutional restricted Boltzmann machines (cRBMs) to model the shape of both tumor core and complete tumor, which includes edema and core. The cRBMs are trained on expert segmentations of training images, without the use of the intensity information in the training images. Experiments on public benchmark data of patients suffering from low- and high-grade gliomas show that the method performs well compared to current state-of-the-art methods, while not being tied to any specific imaging protocol.


Markov Chain Monte Carlo Gaussian Mixture Model Tumor Core Tumor Segmentation Visible Unit 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.



This research was supported by NIH NCRR (P41-RR14075), NIBIB (R01EB013565) and the Lundbeck Foundation (R141-2013-13117).


  1. 1.
    Tustison, N., et al.: ANTs and Árboles. In: Proceedings of the MICCAI-BRATS 2013 (2013)Google Scholar
  2. 2.
    Havaei, M., et al.: A convolutional neural network approach to brain tumor segmentation. In: Proceedings of the MICCAI-BRATS 2015 (2015)Google Scholar
  3. 3.
    Pereira, S., et al.: Deep convolutional neural networks for the segmentation of gliomas in multi-sequence MRI. In: Proceedings of the MICCAI-BRATS 2015 (2015)Google Scholar
  4. 4.
    Menze, B.H., et al.: The multimodal brain tumor image segmentation benchmark (BRATS). IEEE Trans. Med. Imaging 34(10), 1993–2024 (2015)CrossRefGoogle Scholar
  5. 5.
    Kwon, D., Shinohara, R.T., Akbari, H., Davatzikos, C.: Combining generative models for multifocal glioma segmentation and registration. In: Golland, P., Hata, N., Barillot, C., Hornegger, J., Howe, R. (eds.) MICCAI 2014, Part I. LNCS, vol. 8673, pp. 763–770. Springer, Heidelberg (2014)Google Scholar
  6. 6.
    Bakas, S., et al.: Segmentation of gliomas in multimodal magnetic resonance imaging volumes based on a hybrid generative-discriminative framework. In: Proceedings of the MICCAI-BRATS 2015 (2015)Google Scholar
  7. 7.
    Menze, B.H., et al.: Segmenting glioma in multi-modal images using a generative model for brain lesion segmentation. In: Proceedings of the MICCAI-BRATS 2012 (2012)Google Scholar
  8. 8.
    Haeck, T., et al.: Automated model-based segmentation of brain tumors in MR images. In: Proceedings of the MICCAI-BRATS 2015 (2015)Google Scholar
  9. 9.
    Leemput, V., et al.: Automated model-based tissue classification of MR images of the brain. IEEE Trans. Med. Imaging 18(10), 897–908 (1999)CrossRefGoogle Scholar
  10. 10.
    Ashburner, J., et al.: Statistical Parametric Mapping. The Wellcome Department Cognitive Neurology, University College London, London, UK.
  11. 11.
    Murphy, K.P.: Machine learning: a probabilistic perspective. MIT Press, Cambridge (2012)Google Scholar
  12. 12.
    Lee, H., et al.: Unsupervised learning of hierarchical representations with convolutional deep belief networks. Commun. ACM 54(10), 95–103 (2011)CrossRefGoogle Scholar
  13. 13.
    Fischer, A., et al.: Training restricted boltzmann machines: an introduction. Pattern Recogn. 47(1), 25–39 (2014)CrossRefMATHGoogle Scholar
  14. 14.
    Melchior, J., et al.: How to center binary restricted boltzmann machines. arXiv preprint (2013). arXiv:1311.1354
  15. 15.
    Kistler, M., et al.: The virtual skeleton database: an open access repository for biomedical research and collaboration. J. Med. Internet Res 15(11), e245 (2013)CrossRefGoogle Scholar

Copyright information

© Springer International Publishing Switzerland 2016

Authors and Affiliations

  • Mikael Agn
    • 1
  • Oula Puonti
    • 1
  • Per Munck af Rosenschöld
    • 2
  • Ian Law
    • 3
  • Koen Van Leemput
    • 1
    • 4
  1. 1.Department of Applied Mathematics and Computer ScienceTechnical University of DenmarkKongens LyngbyDenmark
  2. 2.Department of Oncology, RigshospitaletCopenhagen University HospitalCopenhagenDenmark
  3. 3.Department of Clinical Physiology, Nuclear Medicine and PET, RigshospitaletCopenhagen University HospitalCopenhagenDenmark
  4. 4.Martinos Center for Biomedical Imaging, MGHHarvard Medical SchoolBostonUSA

Personalised recommendations