Classification of frozen-thawed pork loins based on the freezing conditions and thawing losses using the hyperspectral imaging system

Abstract

This study investigated the suitability of a hyperspectral imaging (HSI) system for the classification of frozen-thawed pork loins according to their quality properties. The pork loin slices were frozen at −20, −50, and −70 °C for 1, 2, and 3 months (the 9 freezing conditions). After thawing pork loins at 2 °C, the hyperspectral image was obtained. The photomicrographs of the loins showed that the extracellular spaces were the biggest in the loins frozen at −20 °C for 3 months. The denaturation of myofibrillar proteins measured by the intrinsic tryptophan intensity and surface hydrophobicity was higher in the loins frozen at −20 °C than that of loins frozen at −50 and −70 °C for 2 and 3 months (P < 0.05). The highest and lowest thawing loss was observed in loins frozen at −20 °C for 3 months (9.1 %) and at −70 °C for 1 month (3.6 %), respectively. The classification by the HSI system for 10-class (the 9 freezing conditions and the 1 fresh loin) showed that the highest correct classification (CC%) rates were 83.20 % and 81.82 % in the calibration and prediction sets, respectively, when partial least squares discriminant analysis (PLS-DA) with pre-processing by baseline offset and second derivative was used. In addition, 93.36 % and 91.92 % of CC in the calibration and prediction sets, respectively, were found in the classification of 4-class (the 3 thawing losses and the 1 fresh loin) with the PLS-DA and read-once-write-many-columnar. This study demonstrates that the HSI system can be used to present information on the quality of frozen-thawed pork loin.