Preventing Chaining in Alpha-Trees Using Gabor Filters

Xiaoxuan Zhang, M.H.F. Wilkinson

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

Abstract

Hierarchical segmentation using α-trees can suffer from unwanted leakage or chaining effects, which lower segmentation quality by reducing the depth of the hierarchy. In this paper we introduce a new way to prevent the chaining effect of α-trees. It relies on the odd 2-D Gabor filter. A series of clean, noisy, and blurred synthetic images was used to test the ability of improved α-trees to stop the chaining effect obtaining a 99.8% segmentation accuracy, and we compared it with the contrast-based α-tree proposed previously by Soille. Two remote sensing images were also used to test the performance of the methods on natural images. The results showed that both 4-CN and 8-CN odd Gabor filter based α-trees can prevent the chaining effect efficiently.
Original languageEnglish
Title of host publicationMathematical Morphology and Its Applications to Signal and Image Processing
EditorsBernhard Burgeth, Andreas Kleefeld, Benoit Naegel, Nicolas Passat, Benjamin Perret
PublisherSpringer
Pages268-280
ISBN (Electronic)978-3-030-20867-7
ISBN (Print)978-3-030-20866-0
DOIs
Publication statusPublished - 8-Jul-2019
EventInternational Symposium on Mathematical Morphology - Saarland University, Saarbrücken, Germany
Duration: 8-Jul-201910-Jul-2019
Conference number: 14
http://ismm.uni-saarland.de/

Publication series

NameLecture Notes in Computer Science
Volume11564

Conference

ConferenceInternational Symposium on Mathematical Morphology
Abbreviated titleISMM
CountryGermany
CitySaarbrücken
Period08/07/201910/07/2019
Internet address

Keywords

  • image segmentation
  • mathematical morphology
  • Alpha-tree
  • Gabor filter

Cite this