Sign in Subscribe
Artificial neural networks

Advances in Networks: Robust Shrimp Disease Detection Using Multi-model Convolutional Neural Networks

Strategic Insights

— 1 min read

Researchers from the Isparta University of Applied Sciences have made significant strides in the development of early and accurate detection methods for viral shrimp diseases. According to a new report, convolutional neural networks (CNNs) have emerged as a promising solution for nondestructive identification of shrimp diseases. However, individual CNN models may have limitations in accurately classifying these diseases. To address this issue, combining the outputs of multiple CNN models using ensemble learning approaches can be advantageous. The study aims to classify shrimp diseases using multiple CNN models and ensemble learning strategies, employing beta normalization, hard voting, and weighted ensemble learning approaches.

Key Takeaways:

  • The study used 11 different pre-trained CNN models, including MobileNet, DenseNet169, DenseNet121, and DenseNet201, to classify shrimp diseases.
  • The MobileNet model achieved the highest individual performance, with an average accuracy of 0.919 +/- 0.001.
  • The weighted learning strategy (WM-3) using the top four models achieved an average accuracy of 0.973 +/- 0.004.
  • The Gradient-weighted Class Activation Mapping (Grad-CAM) method was used to evaluate the decision-making mechanisms of the models.
  • Statistical evaluations were performed using the Wilcoxon Signed-Rank test and Cohen's d effect size analysis.
  • The research concluded that utilizing ensemble strategies with a combination of heterogeneous CNN models can significantly improve the accuracy of shrimp disease classification compared to individual CNN models.
  • The study has been peer-reviewed and published in the Aquacultural Engineering journal.

Statistics:

  • 11 different pre-trained CNN models were used in the study.
  • 5-fold cross-validation was employed to evaluate the performance of the models.
  • The MobileNet model achieved an average accuracy of 0.919 +/- 0.001.
  • The weighted learning strategy (WM-3) using the top four models achieved an average accuracy of 0.973 +/- 0.004.
  • The Wilcoxon Signed-Rank test was used to evaluate the statistical significance of the results.
  • Cohen's d effect size analysis was used to evaluate the magnitude of the effect.

Sources:

  • Robust Shrimp Disease Detection Using Multi-model Convolutional Neural Networks-based Ensemble Strategies. Aquacultural Engineering, 2025;111.
  • NewsRx. New Findings Reported from Isparta University of Applied Sciences Describe Advances in Networks (Robust Shrimp Disease Detection Using Multi-model Convolutional Neural Networks-based Ensemble Strategies). Journal of Engineering. October 20, 2025; p 1752.

Read more