Agglomerative clustering on range data with a unified probabilistic merging function and termination criterion

Steven M Lavalle, Kenneth J. Moroney, Seth Andrew Hutchinson

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

Abstract

Clustering methods, which are frequently employed for region-based segmentation, are inherently metric based. A fundamental problem with an estimation-based criterion is that as the amount of information in a region decreases, the parameter estimates become extremely unreliable and incorrect decisions are likely to be made. We show that clustering need not be metric based, and further we use a rigorous region merging probability function that makes use of all information available in the probability densities of a statistical image model. Also, by using this probability function as a termination criterion, we can produce segmentations in which all region merges were performed above some level of confidence.

Original languageEnglish (US)
Title of host publicationIEEE Computer Vision and Pattern Recognition
Editors Anon
PublisherPubl by IEEE
Pages798-799
Number of pages2
ISBN (Print)0818638826
StatePublished - Dec 1 1993
EventProceedings of the 1993 IEEE Computer Society Conference on Computer Vision and Pattern Recognition - New York, NY, USA
Duration: Jun 15 1993Jun 18 1993

Publication series

NameIEEE Computer Vision and Pattern Recognition

Other

OtherProceedings of the 1993 IEEE Computer Society Conference on Computer Vision and Pattern Recognition
CityNew York, NY, USA
Period6/15/936/18/93

    Fingerprint

ASJC Scopus subject areas

  • Engineering(all)

Cite this

Lavalle, S. M., Moroney, K. J., & Hutchinson, S. A. (1993). Agglomerative clustering on range data with a unified probabilistic merging function and termination criterion. In Anon (Ed.), IEEE Computer Vision and Pattern Recognition (pp. 798-799). (IEEE Computer Vision and Pattern Recognition). Publ by IEEE.