Skip to main navigation Skip to search Skip to main content

Virtually Redying Histological Images with Generative Adversarial Networks to Facilitate Unsupervised Segmentation: A Proof-of-Concept Study

  • M. Gadermayr*
  • , B.M. Klinkhammer
  • , P. Boor
  • *Corresponding author for this work
  • Institute of Imaging & Computer Vision, RWTH Aachen University
  • Institute of Pathology, RWTH Aachen University Hospital

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Approaches relying on adversarial networks facilitate image-to-image-translation based on unpaired training and thereby open new possibilities for special tasks in image analysis. We propose a methodology to improve segmentability of histological images by making use of image-to-image translation. We generate virtual stains and exploit the additional information during segmentation. Specifically a very basic pixel-based segmentation approach is applied in order to focus on the information content available on pixel-level and to avoid any bias which might be introduced by more elaborated techniques. The results of this proof-of-concept trial indicate a performance gain compared to segmentation with the source stain only. Further experiments including more powerful supervised state-of-the-art machine learning approaches and larger evaluation data sets need to follow. © 2019, Springer Nature Switzerland AG.
Original languageEnglish
Title of host publicationDigital Pathology
Subtitle of host publication15th European Congress, ECDP 2019, Warwick, UK, April 10–13, 2019, Proceedings
PublisherSpringer Nature
Pages38-46
Number of pages9
Volume11435 LNCS
ISBN (Electronic)978-3-030-23937-4
ISBN (Print)978-3-030-23936-7
DOIs
Publication statusPublished - 2019
Event15th European Congress on Digital Pathology, ECDP 2019 - Warwick, United Kingdom
Duration: 10 Apr 201913 Apr 2019

Publication series

NameLecture Notes in Computer Science
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference15th European Congress on Digital Pathology, ECDP 2019
Abbreviated titleECDP 2019
Country/TerritoryUnited Kingdom
CityWarwick
Period10/04/1913/04/19

Keywords

  • Adversarial networks
  • Glomeruli
  • Histology
  • Kidney
  • Segmentation
  • Tubuli
  • Unsupervised
  • Image enhancement
  • Pathology
  • Pixels
  • Image segmentation

Fingerprint

Dive into the research topics of 'Virtually Redying Histological Images with Generative Adversarial Networks to Facilitate Unsupervised Segmentation: A Proof-of-Concept Study'. Together they form a unique fingerprint.

Cite this