Abstract
ABSTRACT The prediction of human assigned marbling scores and taste panel evaluations of meat palatability using information about the visual appearance of fat in beef carcass ribeyes was investigated. The distribution of fat within the ribeye was modeled as a realization of a Boolean random set. Model parameters estimated from seventy carcasses, representing six USDA marbling classes, were used as pattems in a hierarchical clustering scheme. Results of the clustering analysis showed that two distinct marbling classes could be distinguished with an acceptable error rate. These results were contrasted with a classification scheme based solely on the area of marbling within the ribeye. The comparison showed that only slight improvements in classification rate were realized by using the more complex model of fat distribution. In addition, very little correspondence was found between the visual appearance of the marbling, as encapsulated in the Boolean model, and taste parameters. It was concluded that the human grader used in the experiments could resolve two levels of marbling, and the differentiation was not significantly influenced by fat distribution.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call