Skip to main navigation Skip to search Skip to main content

A new method of texture binarization

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

4 Citations (Scopus)

Abstract

In the paper a new method of binarization of grey scale regular textures is presented. The approach is based on a model considering the image points as a two dimensional lattice, on which a random walk of a virtual particle is investigated. The model assumes that the image points possess a potential energy equal to their grey scale values and the particle changes its position according to the Gibbs distribution. The described method assigns to each lattice point, the probability that the particle stays in its current position. The elements of in this way obtained matrix are then transformed and treated like the original image, and the same procedure is repeated. The successive iterations lead to a matrix, which is the binary representation of the initial texture. The main advantage of this method is its ability to binarize regular textures with nonuniform brightness.

Original languageEnglish
Title of host publicationComputer Analysis of Images and Patterns - 7th International Conference, CAIP 1997, Proceedings
EditorsGerald Sommer, Kostas Daniilidis, Josef Pauli
PublisherSpringer Verlag
Pages629-636
Number of pages8
ISBN (Print)3540634606, 9783540634607
DOIs
Publication statusPublished - 1997
Event7th International Conference on Computer Analysis of Images and Patterns, CAIP 1997 - Kiel, Germany
Duration: 10 Sept 199712 Sept 1997

Publication series

NameLecture Notes in Computer Science
Volume1296 1296 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference7th International Conference on Computer Analysis of Images and Patterns, CAIP 1997
Country/TerritoryGermany
CityKiel
Period10/09/9712/09/97

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science

Fingerprint

Dive into the research topics of 'A new method of texture binarization'. Together they form a unique fingerprint.

Cite this