Estimation by vision of pork cuts composition for sex and halothane genotype subpopulations

Gérard Daumas et Mathieu Monziols, 63e International Congress of Meat Science and Technology, Cork, Irlande, 13-18 août 2017, p. 52-53

This work aims to quantify the systematic deviations for sex and halothane genotype subpopulations in the estimation by automatic vision of pork cuts composition. A sample of 208 carcasses, startified according to sex (50% castrated males and 50% females), was measured on line by the classification method CSB Image-Meater® (IM). An ear sample was analysed for halothane gene (Hal). The left sides were cut according to the EU procedure and the four main cuts were CT scanned  in order to determine the lean meat percentage (LMP) in each cut. These LMPs were regressed on six IM potential predictors and the carcass weight. For each equation were calculated the systematic deviations per level of the sex and Hal factors by mean difference between predicted and observed values. The absolute deviation per sexual types was the lowest (0.5%) in ham and the highest (1.4%) in belly. The deviations per Hal genotype ranged from 0.3% in shoulder to 0.8% in belly. In all models females and Hal heterozygotes were underestimated and vice-versa. The deviations per Hal genotype were about the half than per sex. The deviations per sex could be removed by at least a different intercept in the prediction equations.