Pattern recognition algorithms

Autores/as

  • Pedro Paulo Leite do Prado Universidade de Taubaté
  • Armando Antonio Monteiro de Castro Universidade de Taubaté

Resumen

This work aims at developing algorithms for pattern recognition, mainly on clustering techniques. Initially, the basic concepts of pattern recognition are presented. Then, a systemic vision of the problem is developed, with the discussion of the approaches for the project of pattern recognition. Finally, the following algorithms for cluster-seeking are implemented: “Maximim Similarity”, “Maximin-Distance”, “K-Means” and “ISODATA”. An user-friendly interface helps to present the programs, written in C++ and in Matlab.

Publicado

2008-10-11

Número

Sección

Artigos