Fish Freshness Detection System Using MobileNetV2
A machine learning web application that classifies fish freshness based on eye and gill images using MobileNetV2, providing fast and accurate predictions through an interactive dashboard.
Overview
Fish Freshness Detection System is a web-based machine learning application developed to assist in determining fish freshness through image classification. The system utilizes the MobileNetV2 deep learning architecture to analyze visual characteristics of fish eyes and gills, which are important indicators of freshness.
Users can upload fish images through an interactive web interface, and the model will classify them into four categories: eye-fresh, eye-non-fresh, gill-fresh, and gill-non-fresh. The application also provides comprehensive model evaluation results, including confusion matrix visualization, precision, recall, F1-score metrics, and overall accuracy.
Built using Python, TensorFlow, and Streamlit, the platform offers a lightweight, user-friendly, and real-time solution for fish quality assessment, helping improve efficiency and consistency in freshness evaluation.
Key Features
- Fish image upload and classification
- Freshness detection based on eye and gill conditions
- Real-time prediction using MobileNetV2
- Confidence score visualization
- Support for JPG, JPEG, and PNG image formats