Assessment of meat freshness with metal oxide sensor microarray electronic nose: A practical approach

We estimate the ability of KAMINA e-nose based on a metal oxide sensor (MOS) microarray and linear discriminant analysis (LDA) pattern recognition to evaluate meat freshness. The received results show that: (1) one or two exposures of standard meat samples to the e-nose are enough for the instrument to recognize with 100% probability the fresh meat prepared by the same supplier; (2) the meat samples of two kinds, stored at 4 and 25 °C, are mutually recognized at early stages of decay with the help of the LDA model built independently under the e-nose training to each kind of meat; (3) the 3–4 training cycles of exposures to meat from different suppliers are necessary for the e-nose to build a reliable LDA model accounting for the supplier factor. This study approves that the MOS e-nose is ready to be currently utilised in food industry for evaluation of product freshness. The e-nose performance is characterized by low training cost, a confident recognition power of various product decay conditions and easy adjustment to changing conditions. In addition, we discuss criteria of e-nose training completeness.