Fish Disease Diagnosis System Based on Image Processing of Pathogens' Microscopic Images

Abstract

This paper presents an application of image processing techniques for rapid disease diagnosis from parasites' microscopic images of diseased fishes. An image processing-based fish disease diagnosis system was developed, allowing for earlier treatment of infected fish to prevent the spread of disease. The system is composed of three steps: 1) extracting pathogen area from the microscopic images of infected fish tissue by applying various kinds of image processing such as noise reduction, edge detection, morphological operations, background extraction, and object detection, 2) confirming pathogen and making final diagnosis by feature extraction, correlation matching and statistical pattern matching compared to the registered pathogen database, and 3) sending a message about diagnosed disease and treatment method to the fish farmers on their mobile phones after sorting out appropriate treatment methods by searching drug information database. Experimental results show that the proposed system is more convenient, consistent, and rapid to diagnosis fish disease rather than general diagnosis method.