Automated seafood freshness detection and preservation analysis using machine learning and paper-based pH sensors.
Journal:
Scientific reports
Published Date:
Jul 18, 2025
Abstract
Seafood, including fish, prawns and various marine products, is a critical component of global nutrition due to its high protein content, essential fatty acids, vitamins and minerals. Traditional methods for assessing seafood freshness such as sensory evaluation and microbiological analysis are labor-intensive, time-consuming and often require specialized equipment. To address these limitations, this research presents an automated freshness detection system for refrigerated fish using machine learning and evaluates the effectiveness of different packaging techniques. Six seafood varieties: Mackerel, Sardine, Prawn, Pomfret, Red Snapper and Cuttlefish were stored under refrigeration and packaged using three methods: vacuum, shrink, and normal packaging. A paper-based pH sensor integrated with Methyl Red and Bromocresol Purple was utilized as a freshness indicator. This sensor effectively monitored spoilage by capturing L*, a*, and b* color values over time. Key parameters, including protein content, lipid content, and Total Volatile Basic Nitrogen (TVB-N) levels, were measured throughout the storage period to assess changes in fish quality. These measurements were used to train a Random Forest (RF) model, aimed at accurately predicting the pH values of the samples. The model's performance was evaluated using standard metrics such as Mean Absolute Error (MAE), Mean Squared Error (MSE), and Root Mean Squared Error (RMSE). Among the tested varieties, Pomfret and Mackerel displayed the lowest MSE values (0.004625 and 0.005034, respectively), indicating high predictive reliability. The RMSE values, representing average prediction error magnitude, were also minimal for Pomfret (0.068007) and Mackerel (0.070949). Furthermore, MAE values confirmed robust predictions, with Pomfret (0.065833) and Mackerel (0.062933) achieving the least deviation from actual measurements. The study demonstrates the effectiveness of the paper-based pH sensor as a visual indicator of spoilage, while the RF-based prediction model offers a reliable method for ensuring food safety and quality during cold-chain storage. Integrating sensor-based monitoring with advanced packaging presents a viable solution for extending the shelf life of seafood and enhancing consumer safety.
