Disease Classification of Oranda Goldfish Using YOLO Object Detection Algorithm
Jeanne Katherine Medina, Pete Jasper P. Tribiana, Jocelyn F. Villaverde
Abstract
Owning an ornamental fish aquarium at home is not easy. It requires much attention to maintain its healthy state. However, if the keeper is a beginner and unable to take care of it properly, a problem may arise. One example is Oranda Goldfish which can easily acquire fish lice, fungi, and white spots. If not recognized at a possible early time, the diseases may spread to the entire aquarium or, worse, the loss of the fish. This study aims to create a system that can diagnose fish disease accurately earlier than the conventional method. The system acquires goldfish images or live feeds using a camera module and undergoes pre-processing to emphasize the essential features. Once the features are segmented, the YOLO algorithm will extract them. The system will then classify any detected disease. The results of the data sampling were able to detect and classify the goldfish samples accurately, with an overall accuracy of 91.4286%. This study utilizing CNN and YOLO helped resolve this problem by diagnosing the common disease of the goldfish. It helps beginner fish farmers, veterinarians, aquarium owners to detect the disease at the earliest time.