Duroc and Iberian pork neural network classification by visible and near infrared reflectance spectroscopy

Abstract

Visible and near infrared reflectance spectroscopy (VIS/NIRS) was used to differentiate between Duroc and Iberian pork in the M. masseter. Samples of Duroc (n = 15) and Iberian (n = 15) pig muscles were scanned in the VIS/NIR region (350–2500 nm) using a portable spectral radiometer. Both mutual information and VIS/NIRS spectra characterization were developed to generate a ranking of variables and the data were then processed by artificial neural networks, establishing 1, 3, or 10 wavelengths as input variable for classifying between the pig breeds. The models correctly classified >70% of all problem assumptions, with a correct classification of >95% for the three-variable assumption using either mutual information ranking or VIS/NIRS spectra characterization. These results demonstrate the potential value of the VIS/NIRS technique as an objective and rapid method for the authentication and identification of Duroc and Iberian pork.