Utilization of Electrical Impedance Spectroscopy and Image Classification for Non-Invasive Early Assessment of Meat Freshness

This study presents a system for assessing the freshness of meat with electrical impedance spectroscopy (EIS) in the frequency range of 125 Hz to 128 kHz combined with an image classifier for non-destructive and low-cost applications. The freshness standard is established by measuring the aerobic pl...

Full description

Saved in:
Bibliographic Details
Published in:Sensors (Basel, Switzerland) Switzerland), 2021-02, Vol.21 (3), p.1001
Main Authors: Huh, Sooin, Kim, Hye-Jin, Lee, Seungah, Cho, Jinwoo, Jang, Aera, Bae, Joonsung
Format: Article
Language:English
Subjects:
Algorithms
Animals
Cattle
Communication
Dielectric Spectroscopy
electrical impedance spectroscopy (EIS)
Food Analysis
freshness evaluation
machine learning
Meat - analysis
Citations: Items that this one cites
Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This study presents a system for assessing the freshness of meat with electrical impedance spectroscopy (EIS) in the frequency range of 125 Hz to 128 kHz combined with an image classifier for non-destructive and low-cost applications. The freshness standard is established by measuring the aerobic plate count (APC), 2-thiobarbituric acid reactive substances (TBARS), and composition analysis (crude fat, crude protein, and moisture) values of the microbiological detection to represent the correlation between EIS and meat freshness. The EIS and images of meat are combined to predict the freshness with the Adaboost classification and gradient boosting regression algorithms. As a result, when the elapsed time of beef storage for 48 h is classified into three classes, the time prediction accuracy is up to 85% compared to prediction accuracy of 56.7% when only images are used without EIS information. Significantly, the relative standard deviation (RSD) of APC and TBARS value predictions with EIS and images datum achieves 0.890 and 0.678, respectively.
ISSN:1424-8220
1424-8220
DOI:10.3390/s21031001