Unsupervised Electromagnetic Target Classification by Self-organizing Map Type Clustering

Katilmis T. T., Ekmekci E., Turhan-Sayan G.

Progress in Electromagnetics Research Symposium, Cambridge, Kanada, 5 - 08 Temmuz 2010, ss.240-243 identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Basıldığı Şehir: Cambridge
  • Basıldığı Ülke: Kanada
  • Sayfa Sayıları: ss.240-243
  • Orta Doğu Teknik Üniversitesi Adresli: Evet

Özet

In this study, design of a completely unsupervised electromagnetic target classifier will be described based on the use of Self-Organizing Map (SOM) type artificial neural network training and Wigner distribution (WD) based target feature extraction technique. The suggested classification method will be demonstrated for a target library of four dielectric spheres which have exactly the same size but slightly different permittivity values.