Pattern Recognition Investigation and Analysis of Planar Point Patterns.

Abstract

Two basic types of planar point patterns were investigated. The first is specified by a set of points in a plane and a noise model which determines the allowed perturbations of the points subject only to the restriction that no interpoint distance should change by more than a certain percentage. The second type is specified by a set of points subject to angle and distance constraints. Representations of point patterns having some invariance properties are presented. For the first type of prototype patterns, these include SIDV's ((Sorted Interpoint Distance Vectors), SSN's (Sorted Nearest Neighbor Vectors), and MST'S (Minimal Spanning Trees). Theorems, experimental results of simulations, advantages, disadvantages and comparisons are presented for classification precedures based on these representations. It is concluded that classification of the first type of patterns can be accomplished using the methods based on SIDV's, SNN's and MST's under certain restrictions. If the number of points in the pattern is small, less than 100, then the SIDV method would be better, while more than 100, then the SNN or methods based on MST's would be more useful. If the percentage of additions or deletions of points is small, of the order of 10%, then the SDV, SNN, and MST are still feasible. If the percentage is large, then other methods must be used.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jan 01, 1979
Accession Number
ADA086766

Entities

People

  • Barbara Lambird
  • David Lavine
  • Laveen N. Kanal

Tags

Communities of Interest

  • Air Platforms
  • C4I
  • Weapons Technologies

DTIC Thesaurus Topics

  • Air Defense
  • Air Force
  • Air Force Facilities
  • Algorithms
  • Cartesian Coordinates
  • Center Of Gravity
  • Classification
  • Coordinate Systems
  • Corporations
  • Counterintelligence
  • Department Of Defense
  • National Security
  • Night Vision
  • Pattern Recognition
  • Real Numbers
  • Recognition
  • Standards

Readers

  • Graph Algorithms and Convex Optimization.
  • Neural Network Machine Learning.
  • Systems Analysis and Design

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference