Sea Turtle Species Classification for Environmental Research and Conservation

Attal Z., DİREKOĞLU C.

10th International Conference on Theory and Application of Soft Computing, Computing with Words and Perceptions (ICSCCW), Prague, Çek Cumhuriyeti, 27 - 28 Ağustos 2019, cilt.1095, ss.580-587, (Tam Metin Bildiri) identifier identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Cilt numarası: 1095
  • Doi Numarası: 10.1007/978-3-030-35249-3_74
  • Basıldığı Şehir: Prague
  • Basıldığı Ülke: Çek Cumhuriyeti
  • Sayfa Sayıları: ss.580-587
  • Orta Doğu Teknik Üniversitesi Kuzey Kıbrıs Kampüsü Adresli: Evet

Özet

We present two different computer vision algorithms for automatic sea turtle species classification. This study effectively contributes to environmental research, and particularly sea turtle conservation projects since sea turtles are endangered species. We classify two different sea turtle species belong to Mediterranean region, namely loggerhead and green sea turtles. The first method we experiment is the Bag of Feature (BoF) method that is integrated with Support Vector Machine (SVM) for classification. The BoF method mainly extracts color, texture and shape information of the sea turtle. The second method we employ is the Convolutional Neural Networks (CNN) that is a very effective algorithm in classification tasks. The dataset used in this work contains more than a thousand of images for each class. Results show that CNN performs maximum 70.14%, while the BoF method performs 69%. Extensive evaluation with respect to different parameter settings is presented in this paper.