Classification of silver halide microcrystals via K-NN clustering of their shape descriptors

Volodymyr V. Kindratenko, Boris A. Treiger, Pierre J.M. Van Espen

Research output: Contribution to journalArticlepeer-review

Abstract

A method for the classification of tabular grain silver halide microcrystals according to their shape is presented. Various approaches of shape analysis and recognition and their applicability for the given problem are discussed. Shape descriptors obtained from Fourier power spectra are used to describe the shape of microcrystals. The classification of the shapes is based on nearest neighborhood algorithms. Results of the classification by four different algorithms are compared. The fuzzy four-nearest-neighbor classifier was found to be the most appropriate one.

Original languageEnglish (US)
Pages (from-to)131-139
Number of pages9
JournalJournal of Chemometrics
Volume11
Issue number2
DOIs
StatePublished - 1997
Externally publishedYes

Keywords

  • Image processing
  • K-NN clustering
  • Shape description

ASJC Scopus subject areas

  • Analytical Chemistry
  • Applied Mathematics

Fingerprint

Dive into the research topics of 'Classification of silver halide microcrystals via K-NN clustering of their shape descriptors'. Together they form a unique fingerprint.

Cite this