Automated seafood freshness detection and preservation analysis using machine learning and paper-based pH sensors

Area of Science:

• Food Science and Technology
• Analytical Chemistry
• Machine Learning Applications

Background:

• Traditional seafood freshness assessment is labor-intensive and time-consuming.
• Developing automated, rapid methods is crucial for global food safety and nutrition.
• Refrigerated seafood quality monitoring requires innovative solutions.

Purpose of the Study:

• To develop and evaluate an automated freshness detection system for refrigerated seafood.
• To assess the effectiveness of different packaging methods (vacuum, shrink, normal) on seafood spoilage.
• To utilize a paper-based pH sensor and machine learning for predicting seafood quality.

Main Methods:

• Six seafood varieties were stored under refrigeration with three packaging types.
• A paper-based pH sensor (Methyl Red, Bromocresol Purple) measured color changes (L*, a*, b*) over time.
• Protein, lipid content, and Total Volatile Basic Nitrogen (TVB-N) were measured; data trained a Random Forest (RF) model to predict pH.

Main Results:

• The paper-based pH sensor effectively indicated spoilage through colorimetric changes.
• The RF model demonstrated high predictive reliability for Pomfret and Mackerel, with low Mean Squared Error (MSE) and Root Mean Squared Error (RMSE).
• MAE values confirmed robust predictions, indicating minimal deviation from actual measurements for these species.

Conclusions:

• The paper-based pH sensor serves as a reliable visual spoilage indicator for seafood.
• The RF-based prediction model offers a robust method for ensuring food safety and quality in cold-chain logistics.
• Integrating sensor technology with advanced packaging is a viable strategy to extend seafood shelf life and improve consumer safety.