Pattern recognition and feature extraction: a comparative study

  • Authors:
  • Vincenzo Niola;Giuseppe Quaremba

  • Affiliations:
  • Department of Mechanical Engineering for Energetics, University of Naples "Federico II", Napoli, Italy;University of Naples "Federico II", Napoli, Italy

  • Venue:
  • NN'05 Proceedings of the 6th WSEAS international conference on Neural networks
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

The selection of features for classifying a pattern by means a fuzzy reasoning, is fundamental in order to obtain a reliable and significative response. The scope of this work is to compare three methods specialized for the extraction of features from images and, consequently, to study the ability of classification performed by applying a fuzzy inference system. The methods to be compared were: Fourier descriptors, Zernike moments and Wavelet coefficients. The best result, in terms of the best performances obtained both as classification reliability and computational time, was represented by the application of wavelet transform.