An Intelligent Scheme of Automatic Fish Disease Recognition Using Optimized Deep Features

Abstract

Aquaculture involves the breeding of aquatic organisms for maintenance and conservation purposes. Detecting fish diseases remains a major challenge in aquaculture because of the diversity of fish species and the influence of environmental factors. This study has implications for improving fish disease detection in aquaculture. In this study, a dataset consisting of healthy and diseased fish images was collected from various sources. Several convolutional neural network models, including VGG16, VGG19, ResNet50, Inception V3, and Xception, were employed for automated feature extraction. Then the extracted features from each model were assessed using different machine learning (ML) classifiers, both with and without feature optimization for fish disease identification. For feature optimization, principal component analysis and linear discriminant analysis (LDA) were investigated. To further improve performance, a soft voting ensemble classifier was utilized by combining the outputs of multiple ML classifiers, thereby taking advantage of their individual strengths. The combination of Xception, LDA, and the voting ensemble achieved the highest accuracy of 98.3% for fish disease detection, and this outcome outperforms existing methods. 

Received: 23 March 2025 | Revised: 20 January 2026 | Accepted: 21 July 2026

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

The data that support the findings of this study are openly available in GitHub at https://github.com/meharaj-emon028/Fish-Dataset, reference number [14]. 

Author Contribution Statement

Md. Meharaj Uddin: Methodology, Data curation, Writing – original draft. Ummatun Nahar Oishi: Methodology, Data curation, Writing – original draft. Arslan Munir: Conceptualization, Validation, Investigation, Resources, Writing – review & editing, Supervision. Muhammad Minoar Hossain: Conceptualization, Formal analysis, Writing – review & editing, Project administration. Md. Khabir Uddin Ahamed: Software, Investigation, Visualization. Md. Sadiq Iqbal: Software, Resources.