Multispectral imaging as a predictive tool for freshness of whole Atlantic cod: Compared with sensory, chemical and microbiological analysis

This study explores the potential of using multispectral imaging (MSI) techniques to predict the freshness of whole gutted Atlantic cod (Gadus morhua) throughout its shelf life during storage on ice. Spectral data were acquired from key anatomical regions - the gills, skin, and eyes - and analyzed using chemometrics methods, including partial least squares regression (PLSR) and artificial neural networks (ANNs). These models were trained to predict sensory evaluations performed by trained panelists using the Quality Index Method (QIM) as well as chemical- and microbiological analyses, total viable counts (TVC) and total volatile base nitrogen (TVB-N). Among the regions analyzed, the gills provided the most accurate predictions of the QIM score, with the ANN model achieving an R²_CV = 0.87 and an RMSECV of 2.0. Spectral analysis highlights the role of near-infrared (NIR) wavelengths in capturing spoilage-related biochemical and structural changes, complementing the visible spectrum, which primarily captures color changes. Our findings suggest that MSI combined with chemometric techniques could serve as an efficient, non-destructive alternative to traditional sensory freshness evaluations.