Skip to content
SDGs Center UDINUS

Evidence record · UDINUS · SDGs

Optimization of Ensemble Deep Learning Models for Freshwater Fish Disease Classification Using Particle Swarm Optimization

Rektorat UDINUS 2025 Published 2026-10-30

Evidence record · UDINUS · SDGs

Unit
Rektorat UDINUS
Period
2025
Goals
Record
74916754-a8eb-4fd1-ad9a-f4be941ce7c1
Published
2026-10-30

Universitas Dian Nuswantoro conducted research on the optimization of ensemble deep learning models for classifying diseases in freshwater fish. The activity addressed disease outbreaks in aquaculture, which the source identifies as threats to food security and economic stability, particularly in freshwater fish farming. The research responded to limitations in traditional diagnostic methods, which are described as time-consuming and inefficient, by developing an intelligent and automated classification approach.

The researchers from Universitas Dian Nuswantoro, Maarif, Mochammad Rafli Samsul, Shidik, Guruh Fajar, and Purwanto, integrated four pre-trained architectures: ResNet-50, DenseNet-121, InceptionV3, and EfficientNetB3. Particle Swarm Optimization was used to dynamically adjust the model weights based on performance metrics. A comprehensive data augmentation pipeline expanded a small initial dataset into a balanced collection of 2,100 images representing seven disease classes. The activity also applied Grad-CAM as an Explainable AI technique to improve the interpretability of the classification results.

Experimental results reported that the optimized ensemble achieved 98% accuracy and balanced F1-scores. The optimized ensemble outperformed both individual models and unoptimized ensembles. The results were statistically validated using paired t-tests. The source presents the approach as robust, interpretable, and scalable for automated freshwater fish disease diagnosis and connects it with smarter and more sustainable aquaculture practices. In relation to the stated goals, the activity concerns food security, responsible and sustainable aquaculture practices, and the protection of freshwater-related biological resources through disease classification research.

Reference https://www.scopus.com/pages/publications/105036856434?origin=resultslist