Scene Classification Results Using the Max-Min Texture Measure.

Abstract

Results from a previous analysis by ETL indicated that the Max-Min texture measure was practical and showed promise of being an efficient tool for image segmentation. Probabilistic relaxation was applied to the results in order to remove ambiguities and to remove noise. Results from the relaxation process were also encouraging and it was decided that both processes, i.e., the classification using Max-Min texture followed by a relaxation of the output, would be applied to full scenes to determine if the techniques were still applicable. Note that tests in the referenced experiment were limited to regions selected to train the process. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1982
Accession Number
ADA123496

Entities

People

  • Michael A. Crombie
  • Nancy J. Friend
  • Robert S. Rand

Organizations

  • Geospatial Research Laboratory

Tags

Communities of Interest

  • Materials and Manufacturing Processes

DTIC Thesaurus Topics

  • Army
  • Army Corps Of Engineers
  • Boundaries
  • Classification
  • Computer Vision
  • Computers
  • Data Science
  • Databases
  • Engineers
  • Image Segmentation
  • Information Processing
  • Information Science
  • Machine Learning
  • Neurobehavioral Manifestations
  • Sampling
  • Security

Readers

  • Computer Vision.
  • Systems Analysis and Design