Two-Stage Fish Disease Diagnosis System Based on Clinical Signs and Microscopic Images

Abstract

This paper presents a two-stage, fish disease diagnosis system capable of faster treatment of diseased fish in order to prevent the spread of disease. The two stages are a clinical sign-based diagnosis from the initial sketchy observations of the symptoms and a more thorough microscopic image-based diagnosis from pathogenic detection using image processing techniques. In the first stage, the system suggests candidate diseases with parsed selection based on water temperature, growth phase of the diseased fish, external clinical signs, internal clinical signs and microscopic observations. In the second stage, if the system in the first stage previously suggested using microscopic diagnosis for final diagnosis, the system determines the final disease by discriminating potential pathogens from microscopic images using image pattern recognition techniques and provides a suitable treatment method and guidance in the use of appropriate drugs.

The designed fish disease diagnosis system was implemented to diagnose 14 diseases of olive flounder in the first stage and 3 parasitic diseases in the second stage. The information on diagnosed disease, treatment and prevention methods was provided by a connected web server through internet and SMS message by mobile communication. The system can support fish farmers and veterinarians by providing easy and rapid diagnosis of diseased olive flounder, guidance in the use of appropriate drugs and a suitable treatment method for the diagnosed disease.