Identification of yellowfin tuna (Thunnus albacares), mackerel tuna (Euthynnus affinis), and skipjack tuna (Katsuwonus pelamis) using deep learning

Abstract Yellowfin tuna (Thunnus albacares), mackerel tuna (Euthynnus affinis), and skipjack tuna (Katsuwonus pelamis) have important economic values for the capture fisheries in Indonesia. Activities of identifying these fish and other types of tuna have been done manually, which can lead to errors and ultimately affect statistics, stock estimates, or traceability. The aim of this research is to use deep learning methods in identifying three species of tuna, specifically yellowfin tuna, mackerel tuna, and skipjack tuna. YOLO’s newest model, YOLOv5, was used to identify the fish. The number of epochs that produces the optimum accuracy value for use in the YOLOv5 model is 400. The values for training loss, accuracy, precision, recall and F1-Score when the model is learning with a total of 400 epochs are 0.000253, 95%, 98.1%, 93.9%, and 96%. Based on these results, the three species of tuna can be identified with high accuracy.