Scale-Based Robust Image Segmentation

Abstract

Image segmentation represents an essential step in the early stages of an Automatic Target Recognition system. We propose two robust approaches fundamentally based on scale information inherent to a given imagery. The first approach is parametric in that the scale evolution of an image is statistically captured by a model which is in turn utilized to classify the pixels in the image. The second, on the other hand, non-linearly evolves the image along some specific characteristic to unravel and delineate the various comprising entities in it.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 1997
Accession Number
ADA457838

Entities

People

  • A. Kim
  • A. S. Willsky
  • Hamid Krim
  • I. Pollak

Organizations

  • Massachusetts Institute of Technology

Tags

Communities of Interest

  • Human Systems
  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Boundaries
  • Classification
  • Computer Vision
  • Detectors
  • Differential Equations
  • Hypotheses
  • Identification
  • Image Segmentation
  • Models
  • Multiscale Models
  • Personal Information Managers
  • Probability
  • Probability Distributions
  • Recognition
  • Statistics
  • Synthetic Aperture Radar

Fields of Study

  • Computer science

Readers

  • Computer Vision.
  • Systems Analysis and Design
  • Theoretical Analysis.