Siamese Content Loss Networks for Highly Imbalanced Medical Image - - PowerPoint PPT Presentation

▶
siamese content loss networks for highly imbalanced
SMART_READER_LITE
LIVE PREVIEW

Siamese Content Loss Networks for Highly Imbalanced Medical Image - - PowerPoint PPT Presentation

Feb 20, 2023 306 likes •438 views

Siamese Content Loss Networks for Highly Imbalanced Medical Image Segmentation Brandon Mac 1 , Alan R. Moody 2,3 , April Khademi 1,4 . 1 Image Analysis in Medicine Lab (IAMLAB), Ryerson University 2 Department of Medical Imaging, University of

slide-1
SLIDE 1

Siamese Content Loss Networks for Highly Imbalanced Medical Image Segmentation

Brandon Mac1, Alan R. Moody2,3, April Khademi1,4.

1Image Analysis in Medicine Lab (IAMLAB), Ryerson University 2Department of Medical Imaging, University of Toronto 3Department of Medical Imaging, Sunnybrook Health Sciences Centre 4Keenan Research Centre for Biomedical Science, St. Michael’s Hospital
slide-2
SLIDE 2

Clinical Problem

  • White Matter Hyperintensities [1]

➢Local macroscopic tissue structure erosion ➢Increased water content due to inflammation

  • Key Biomarker

➢Linked with stroke and dementia

  • Typical analysis is manual

➢Time consuming and expensive ➢High inter- and intra- rater variability ➢Reported DSC of 0.66 [2] – 0.83 [3] between radiologists

  • Our Method

➢Trained on: MICCAI 2017 WMH Grand Challenge Data (60 Volumes) ➢Validated on: Canadian Atherosclerosis Imaging Network (50 Volumes)

slide-3
SLIDE 3

Class Imbalance Across Subjects

0.2 0.4 0.6 0.8 1 1.2

Postive Pixel Ratio (%) Subject

Positive Pixel Ratio Across Subjects

slide-4
SLIDE 4

Method – Overview

slide-5
SLIDE 5

Method – Generator

slide-6
SLIDE 6

Method – Discriminator

slide-7
SLIDE 7

Results – Segmentation

slide-8
SLIDE 8

Results – Dice Scores

slide-9
SLIDE 9

Results – Average Volume Difference

slide-10
SLIDE 10

Future Works

  • Explore configurations of the discriminator network
  • Explore GANs and Deep Metric Learning methods

Min Max

slide-11
SLIDE 11

Thank you for Watching

Brandon Mac IAMLAB Ryerson University bmac@ryerson.ca @BrandNameTensor Alan A. Moody University of Toronto Sunnybrook Health Sciences Centre alan.moody@sunnybrook.ca April Khademi IAMLAB Ryerson University

  • St. Michael’s Hospital

akhademi@ryerson.ca @aprilkhademi

slide-12
SLIDE 12

References

[1] H. J. Kuijf et al, "Standardized Assessment of Automatic Segmentation of White Matter Hyperintensities; Results of the WMH Segmentation Challenge," IEEE Transactions on Medical Imaging, vol. 38, (11), pp. 1-1, 2019. [2] C. Egger et al, "MRI FLAIR lesion segmentation in multiple sclerosis: Does automated segmentation hold up with manual annotation?" NeuroImage Clinical, vol. 13, (C), pp. 264-270, 2017.[3] T. Falk et al, "U-Net: deep learning for cell counting, detection, and morphometry," Nature Methods, vol. 16, (1),

  • pp. 67-70, 2019.

[3] M. D. Steenwijk et al, "Accurate white matter lesion segmentation by k nearest neighbor classification with tissue type priors (kNN- TTPs)," NeuroImage. Clinical, vol. 3, pp. 462-469, 2013.