An improved deep CNN-based freshwater fish classification with cascaded bio-inspired networks
Bangladesh has plentiful water, which is essential to its freshwater fish traditions. Environmental concerns and other causes have reduced the country's water resources, threatening many native freshwater fish species. Thus, the younger generation deficiencies recognition of local freshwater fi...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Taylor & Francis Group
2025-04-01
|
| Series: | Automatika |
| Online Access: | https://www.tandfonline.com/doi/10.1080/00051144.2025.2457803 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Bangladesh has plentiful water, which is essential to its freshwater fish traditions. Environmental concerns and other causes have reduced the country's water resources, threatening many native freshwater fish species. Thus, the younger generation deficiencies recognition of local freshwater fish and struggles to recognize them. Traditional methods are very insufficient to overcome these issues. To address these research gaps, the research proposes an automatic system for categorizing Bangladesh's freshwater fish. The proposed methodology involves several key steps, including building a comprehensive dataset, extracting features from fish images using pre-trained Convolutional Neural Network (CNN) models, and employing typical ML approaches. Initially comprising eight classes, the dataset undergoes feature extraction using CNN algorithms, followed by the utilization of various feature selection methods such as Support Vector Classifier, Principal Component Analysis, Linear Discriminant Analysis, and optimization models like Particle Swarm Optimization, Bacterial Foraging Optimization, and Cat Swarm Optimization. In the final phase, seven conventional ML techniques are applied to classify the images of local fishes. Empirical measurements are gathered and analyzed to assess the proposed framework's performance. Particularly, the present approach achieves the highest accuracy of 98.71% through the utilization of the ML mechanism Logistic Regression with Resnet50, SVC, and CSO models. |
|---|---|
| ISSN: | 0005-1144 1848-3380 |