Multi-scale curvatures for contour feature extraction

Koichiro Deguchi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)


A method is described to extract features of contour of a figure based on its curvature. For a digital figure, curvatures are approximated by using differences of contour tangential angles at two points apart from each other with an appropriate distance scale. In this method, a set of curvatures with various scales is calculated, and from the curvature changes with respect to the multiple scales, the features of the original contour, such as the locations and angles of vertices, are extracted. The view-field resolutions for the analysis of features of figures are determined automatically, for each figure and its parts.

Original languageEnglish
Title of host publicationProceedings - International Conference on Pattern Recognition
PublisherPubl by IEEE
Number of pages3
ISBN (Print)0818608781
Publication statusPublished - 1988 Dec 1

Publication series

NameProceedings - International Conference on Pattern Recognition

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition


Dive into the research topics of 'Multi-scale curvatures for contour feature extraction'. Together they form a unique fingerprint.

Cite this