Objective carcass grading for bovine animals based on carcass length

Abstract

ABSTRACT The aim of the study was to evaluate performance for beef carcass grading using a novel semi-automatic method compared to human classifiers. The novelty was measuring carcass length as a predictor. The grading is given as conformation and fat cover as defined by the EUROP classification system. A training set was used to fit the model with predictors based on weight, age, breed and sex, in addition to length. Prediction performance was evaluated for a test set including carcasses graded by Norwegian classifiers, and a separate test set for carcasses graded by international classifiers. The precision for conformation was high (Pearson correlation ≥ 0.94) for both test sets, but the precision for fat cover was lower (Pearson correlation range 0.30–0.91). High correlation for conformation, together with low bias estimates, provides indication that the objective method is equipped to replace the previous human classifier system for conformation in Norway.