Meat and fish freshness inspection system based on odor sensing

Area of Science:

• Food Science
• Sensor Technology
• Machine Learning

Background:

• Contaminated meat poses a significant food safety risk.
• Traditional methods for detecting spoiled meat can be time-consuming and subjective.
• There is a need for rapid and objective detection methods in retail environments.

Purpose of the Study:

• To develop a simple electronic nose for identifying spoiled meat in butcher shops.
• To evaluate the performance of the electronic nose using different pattern classification algorithms.
• To determine the most effective algorithm for spoiled meat detection.

Main Methods:

• Utilized a metal oxide semiconductor-based electronic nose.
• Measured smell signatures from fresh and decayed beef and fish samples.
• Applied artificial neural network, support vector machine, and k-nearest neighbor algorithms for classification.

Main Results:

• The electronic nose successfully differentiated between fresh and spoiled meat samples.
• The k-nearest neighbor algorithm achieved the highest accuracy in identifying decayed meat.
• Performance was evaluated based on accuracy, sensitivity, and specificity.

Conclusions:

• A simple electronic nose can effectively detect spoiled meat using commercial sensors.
• The k-nearest neighbor algorithm is a promising tool for spoiled meat identification.
• This technology offers a potential solution for enhancing food safety in retail settings.