Non-destructively Prediction of Quality Parameters of Dry-Cured Iberian Ham by Applying Computer Vision and Low-Field MRI

Computer vision algorithms and Magnetic Resonance Imaging (MRI) have been proposed to obtain quality traits of Iberian hams, due to the non-destructive, non-ionizing and innocuous nature of these approaches. However, all the proposals have been based on high-field MRI scanners, which obtain high quality images but also involve very high economical costs. In this paper, low-field MRI devices and three classical texture algorithms were used to predict quality traits of Iberian ham. Prediction equation of quality features were obtained, which estimate the quality parameters as a function of computational textures. The texture features were obtained by applying three well-known classical texture algorithms (GLCM - Gray Level Co-occurrence Matrix, GLRLM - Gray Level Run Length Matrix and NGLDM - Neighbouring Gray Level Dependence Matrix) on low-field MRI. Being the first approach that exploits this type of scanner for this purpose in dry-cured meat, the predicted elements were compared and correlated to the results obtained by means of traditional physico-chemical methods. The obtained correlation were higher than 0.7 for almost all the quality traits, reached very good to excellent relationship. These high correlations between both sets of data (traditional and estimated results) prove that low-field MRI combined with texture algorithms could be used to estimate the quality traits of meat products in a non-destructive and efficient way.