Determination of Fish Health Using Machine Learning and Image Processing Techniques

Creative Commons License

Çira D., Can E., Austin B.

9. Uluslararası Öğrenci Bilim Kongresi (ISSC)/9. International Students Science Congress (ISSC), Manisa, Türkiye, 22 - 23 Mayıs 2025, ss.89, (Özet Bildiri)

  • Yayın Türü: Bildiri / Özet Bildiri
  • Basıldığı Şehir: Manisa
  • Basıldığı Ülke: Türkiye
  • Sayfa Sayıları: ss.89
  • İstanbul Üniversitesi-Cerrahpaşa Adresli: Evet

Özet

In aquaculture, managing the health of fish is of great importance for effective production. Detecting individuals showing disease signs in the production system is very important, involving onsite clinical and laboratory-based examinations by veterinarians and fish health professionals. The aim is that disease diagnosis is made correctly, allowing effective treatment regimes to be implemented. However, identifying the presence of diseased individuals in the stock may be difficult, especially in intensively reared populations. Effective disease diagnosis and control may be time consuming with inevitable serious economic losses. In recent years, machine learning and image processing techniques have become important for monitoring animal populations. In aquaculture, image processing techniques are also used to evaluate the health and welfare status of fish by analyzing their fin structure, color changes, body shape and behavioral abnormalities. Artificial intelligence-supported systems can automatically detect disease symptoms by processing data obtained from underwater cameras or thermal imaging devices. Conversely, machine learning algorithms achieve high accuracy rates in distinguishing healthy and diseased fish by learning from large data sets. Techniques, such as support vector machines, deep learning models and decision trees, provide early diagnosis by analyzing the body temperature, swimming behavior and gill movements of fish. With these technologies, diseased fish may be detected at an early stage, enabling appropriate control measures to be instigated quickly, and thus minimizing losses. Therefore, these new technologies lead to increased production efficiency and sustainable farming. As a result, machine learning and image processing techniques offer innovative and effective solutions for sustainable aquaculture. Future research should focus on more species-specific and detailed studies.

Keywords: Machine learning, image processing techniques, fish health, aquaculture, disease diagnosis, deep learning