A morphological signature transform for shape description

Sven Loncaric, Atam P. Dhawan

Research output: Contribution to journalArticlepeer-review

16 Scopus citations

Abstract

A novel shape description method based on the multiresolution morphological image processing by non-convex multiple structuring elements is presented. A binary image which contains the object shape to be described is represented by means of a multiresolution pyramid. A set of primary structuring elements along with their rotated versions is used. The method is based on the successive morphological erosions of the input image at different resolutions by primary and rotated structuring elements. The areas of successively eroded images are computed for each structuring element at each pyramid level. The obtained set of numbers is arranged into vectors, ordered, and used as a shape descriptor. Results are included to show that the method is robust against noise and invariant to both rotation and scale change.

Original languageEnglish (US)
Pages (from-to)1029-1037
Number of pages9
JournalPattern Recognition
Volume26
Issue number7
DOIs
StatePublished - Jul 1993
Externally publishedYes

All Science Journal Classification (ASJC) codes

  • Software
  • Signal Processing
  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Keywords

  • Mathematical morphology
  • Multiresolution pyramid
  • Multiscale representation
  • Shape description
  • Shape matching

Fingerprint Dive into the research topics of 'A morphological signature transform for shape description'. Together they form a unique fingerprint.

Cite this