Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Identification of Salmonella Using TF-module Electronic Nose System

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  Paper number  MBSK 06-207,  ASABE/CSBE North Central Intersectional Meeting . (doi: 10.13031/2013.22367) @2006
Authors:   Szymon Woznica, Suranjan Panigrahi, Catherine Logue, Julie Sherwood, Curt Doetkott, Martin Marchello
Keywords:   Electronic Nose, Intelligent sensors, Artificial Neural Networks, Salmonella contamination of meat, Classification, Food quality

Beef is a highly consumed meat throughout the world, especially in the United States. However, improper handling and storage of beef can lead to its contamination by dangerous pathogenic microorganisms such as Salmonella typhimurium. Volatile metabolites that are emitted by these microorganisms can be detected using an electronic nose. Vacuum packed beef strip loins were procured for the study and were repackaged on polystyrene trays over wrapped with food grade cling film. The TF-module electronic nose system which consists of four thin film metal oxide gas sensors was used to analyze the headspace of these tray packed beef strip loins both in the control samples and also when inoculated with Salmonella typhimurium. The storage study was carried out at 20C for 4 days. Different processing methods were initially used to extract four different features from each sensor signal. Principal component analysis was carried out to reduce the number of features and to find new features which best represent the sensor signals. Classification models were then developed using radial basis neural network and statistical techniques such as LDA and QDA. These models were further validated using leave-one-out and bootstrapping methods. The models developed using the extracted features were able to classify the meat samples into two groups, contaminated" (Salmonella counts < 0.7 log10 cfu/g) and "non-contaminated" (Salmonella counts = 0.7 log10 cfu/g). Using these models, a maximum classification accuracy of 93.8% was obtained.

(Download PDF)    (Export to EndNotes)