Advertisement

Degenerating U-Net on Retinal Vessel Segmentation

What Do We Really Need?
  • Weilin FuEmail author
  • Katharina Breininger
  • Zhaoya Pan
  • Andreas Maier
Conference paper
  • 48 Downloads
Part of the Informatik aktuell book series (INFORMAT)

Zusammenfassung

Retinal vessel segmentation is an essential step for fundus image analysis. With the recent advances of deep learning technologies, many convolutional neural networks have been applied in this field, including the successful U-Net. In this work, we firstly modify the U-Net with functional blocks aiming to pursue higher performance. The absence of the expected performance boost then lead us to dig into the opposite direction of shrinking the U-Net and exploring the extreme conditions such that its segmentation performance is maintained. Experiment series to simplify the network structure, reduce the network size and restrict the training conditions are designed. Results show that for retinal vessel segmentation on DRIVE database, U-Net does not degenerate until surprisingly acute conditions: one level, one filter in convolutional layers, and one training sample. This experimental discovery is both counter-intuitive and worthwhile. Not only are the extremes of the U-Net explored on a well-studied application, but also one intriguing warning is raised for the research methodology which seeks for marginal performance enhancement regardless of the resource cost.

Preview

Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.

Literatur

  1. 1.
    Maier A, Syben C, Lasser T, et al. A gentle introduction to deep learning in medical image processing. Zeitschrift für Medizinische Physik. 2019;.Google Scholar
  2. 2.
    Fu H, Xu Y, Wong DWK, et al. Retinal vessel segmentation via deep learning network and fully-connected conditional random fields. In: ISBI; 2016. .Google Scholar
  3. 3.
    Fu W, Breininger K, Schaffert R, et al. A divide-and-conquer approach towards understanding deep networks. MICCAI. 2019;.Google Scholar
  4. 4.
    Ronneberger O, Fischer P, Brox T. U-Net: convolutional networks for biomedical image segmentation. In: MICCAI; 2015. .Google Scholar
  5. 5.
    Isensee F, Petersen J, Klein A, et al. NnU-Net: self-adapting framework for u-netbased medical image segmentation. arXiv:180910486. 2018;.
  6. 6.
    Zhou Z, Siddiquee MMR, et al. Unet++: a nested u-net architecture for medical image segmentation. In: DLMIA; 2018. .Google Scholar
  7. 7.
    Huang G, Liu Z, Van Der Maaten L, et al. Densely connected convolutional networks. In: CVPR; 2017. .Google Scholar
  8. 8.
    He K, Zhang X, Ren S, et al. Deep residual learning for image recognition. In: CVPR; 2016. .Google Scholar
  9. 9.
    Lin TY, Goyal P, Girshick R, et al. Focal loss for dense object detection. In: Proc IEEE Int Conf Comput Vis; 2017. .Google Scholar
  10. 10.
    Kingma DP. Adam: A method for stochastic optimization. arXiv:14126980. 2014;.

Copyright information

© Springer Fachmedien Wiesbaden GmbH, ein Teil von Springer Nature 2020

Authors and Affiliations

  • Weilin Fu
    • 1
    • 2
    Email author
  • Katharina Breininger
    • 1
  • Zhaoya Pan
    • 1
  • Andreas Maier
    • 1
    • 3
  1. 1.Pattern Recognition LabFriedrich-Alexander UniverstiyErlangen–NürnbergDeutschland
  2. 2.International Max Planck Research School for Physics of Light (IMPRS-PL)DresdenDeutschland
  3. 3.Erlangen Graduate School in Advanced Optical Technologies (SAOT)ErlangenDeutschland

Personalised recommendations