Ship Classification Based On Gradient And Support Vector Machines

Creative Commons License

Hanbay K., Özdemir T. B.

International Conference on Artificial Intelligence and Software Engineering (ICAISE-2024), Salzburg, Avusturya, 6 - 07 Mayıs 2024, ss.1-12, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: Salzburg
  • Basıldığı Ülke: Avusturya
  • Sayfa Sayıları: ss.1-12
  • İnönü Üniversitesi Adresli: Evet

Özet

In this paper, an approach for ship classification based on gradient calculation and gray level co -
occurrence matrix is proposed. Four statistical feature are calculated from gradient magnitude
of SAR image. Thus, the e ffe ct of both differential and textural feature extraction strategy have
been analyzed. The distinctive characteristics of the background and ship regions have been
obtained by the applied hybrid feature extraction scheme. Obtained feature vectors are fed to
the support vector machine for ship classification. Thus, it is observed that the efficiency of the
support vector machine was improved. Experimental results show that the proposed feature
fusion strategy using gradient and statistical calculations can be used fo r ship classification.