Automatic marbling prediction of sliced dry-cured ham using image segmentation, texture analysis and regression

Dry-cured ham is a traditional Mediterranean meat product consumed throughout the world. This productis very variable in terms of composition and quality. Consumer's acceptability of this product is influencedby different factors, in particular, visual intramuscular fat and its distribution across the slice, also known asmarbling. On-line marbling assessment is of great interest for the industry for classification purposes. However,until now this assessment has been traditionally carried out by panels of experts and this methodology cannotbe implement in industry. We propose a complete automatic system to predict marbling degree of dry-curedham slices, which combines: (1) the color texture features of regions of interest (ROIs) extracted automaticallyfor each muscle; and (2) machine learning models to predict the marbling. For the ROIs extraction algorithmmore than the 90% of pixels of the ROI fall into the true muscle. The proposed system achieves a correlationof 0.92 using the support vector regression and a set of color texture features including statistics of eachchannel of RGB color image and Haralick's coefficients of its gray-level version. The mean absolute error was0.46, which is lower than the standard desviation (0.5) of the marbling scores evaluated by experts. This highaccuracy in the marbling prediction for sliced dry-cured ham would allow to deploy its application in thedry-cured ham industry.

Automated summary

Dry-cured ham is a popular Mediterranean meat product consumed worldwide. A complete automatic system combining color texture features and machine learning models has been proposed to predict the marbling degree of dry-cured ham slices. The system achieved a high accuracy in marbling prediction with a correlation of 0.92 and a mean absolute error of 0.46. This system has significant implications for the dry-cured ham industry.