Complexity Analysis and u-net Based Segmentation of Meningeal Lymphatic Vessels

Nazia Tabassum, Michael Ferguson, Jasmin Herz, Scott T. Acton

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

Abstract

This paper showcases a u-net based architecture for the automated segmentation images of meningeal lymphatic vessels. These lymphatic vessels surround the cerebral cortex and have been recently found to drain waste from the brain. Studies on mice have shown loss of memory and impairment in cognitive ability if the vessels' draining capacity does not function adequately. As the meningeal lymphatic vasculature itself is a recent discovery, there is no software tailored for automatically segmenting these images. Instead, segmentation must be performed by hand, which is a tedious and errorprone process. By building an automatic segmentation tool for these vessels, we can provide informatics for understanding and researching them, in a quick and reliable way. A convolutional neural network, called u-net, is adapted to the vessel segmentation application, with the goal of teaching the network how to segment the vessels. Segmentation using u-net is compared to traditional non-learning based segmentation methods using Dice coefficient. Three complexity measures are also proposed to evaluate the segmentation quality: vessel ramification index, porosity, and vessel length. The existence of a technique and associated software to automatically segment and analyze these vessels will drastically speed up subsequent neuroscience research in the field.

Original languageEnglish
Title of host publicationConference Record of the 54th Asilomar Conference on Signals, Systems and Computers, ACSSC 2020
EditorsMichael B. Matthews
PublisherIEEE Computer Society
Pages210-214
Number of pages5
ISBN (Electronic)9780738131269
DOIs
StatePublished - Nov 1 2020
Event54th Asilomar Conference on Signals, Systems and Computers, ACSSC 2020 - Pacific Grove, United States
Duration: Nov 1 2020Nov 5 2020

Publication series

NameConference Record - Asilomar Conference on Signals, Systems and Computers
Volume2020-November
ISSN (Print)1058-6393

Conference

Conference54th Asilomar Conference on Signals, Systems and Computers, ACSSC 2020
Country/TerritoryUnited States
CityPacific Grove
Period11/1/2011/5/20

Keywords

  • complexity
  • learning
  • lymphatics
  • porosity
  • ramification
  • segmentation
  • u-net
  • vasculature

Fingerprint

Dive into the research topics of 'Complexity Analysis and u-net Based Segmentation of Meningeal Lymphatic Vessels'. Together they form a unique fingerprint.

Cite this