Branch Profiles for Shape Analysis

Zayed M. Asiri, Brianna L. Martin, Murk J. Bottema

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

1 Citation (Scopus)

Abstract

Ahstract-The necessity for characterising highly irregularly shaped objects appears in many circumstances, most prominently in biology and medicine, but also in physical sciences and elsewhere. Here, a multi-scale method for quantifying the level of branching in irregular structures is presented to extend the repertoire descriptors of shape. The method was used to classify strains of yeast colonies and to demonstrate differences in structure of newly formed cancellous bone in rats under various experimental conditions. Yeast colonies were classified with an accuracy of 1.000 (n = 10) and classification of newly formed cancellous bone into three classes achieved mean accuracy of 0.853 ±. 088 over 10 runs with data randomly sampled from the same 15 rats each run.

Original languageEnglish
Title of host publication2020 Digital Image Computing
Subtitle of host publicationTechniques and Applications, DICTA 2020
Place of PublicationMelbourne, Australia
PublisherInstitute of Electrical and Electronics Engineers
Number of pages7
ISBN (Electronic)9781728191089
DOIs
Publication statusPublished - 29 Nov 2020
Event2020 Digital Image Computing: Techniques and Applications, DICTA 2020 - Melbourne, Australia
Duration: 29 Nov 20202 Dec 2020

Publication series

Name2020 Digital Image Computing: Techniques and Applications, DICTA 2020

Conference

Conference2020 Digital Image Computing: Techniques and Applications, DICTA 2020
Country/TerritoryAustralia
CityMelbourne
Period29/11/202/12/20

Keywords

  • Shape analysis
  • branch profiles
  • irregularly shaped objects
  • irregular structures
  • branching
  • cancellous bone
  • yeast colonies

Cite this