Segmentation Using Locally Optimal Piecewise Approximations.

Abstract

This note deals with a method of segmenting one-dimensional patterns. It is based on a scheme for detecting natural 'sides' (= maximal intervals of approximately constant slope) on a simple closed curve. Analogously, the method described in this note detects natural 'plateaus' (= maximal intervals of approximately constant value) in a function of one variable. It is shown that these plateaus constitute useful building blocks for segmenting the pattern formed by the function values.

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 1975
Accession Number
ADA012959

Entities

People

  • Andrew F. Blumenthal
  • Azriel Rosenfeld
  • Larry S. Davis

Organizations

  • University of Maryland

Tags

DTIC Thesaurus Topics

  • Applied Computer Science
  • Artificial Intelligence
  • Artificial Intelligence Computing
  • Computational Processes
  • Computer Science
  • Computer Vision
  • Computing-Related Activities
  • Intervals
  • Plateaus

Fields of Study

  • Mathematics

Readers

  • Approximation Theory.
  • Mathematical Modeling and Probability Theory.
  • Speech Processing/Speech Recognition.