0
The objective of the present project was to study the evolution of carcass lean meat percentage and ham composition in growing pigs by means of Hounsfield value distributions from images obtained in live pigs with computed tomography (CT). For this purpose, 60 pigs from 3 genetic lines were anesthetized and CT scanned. The total volume of carcass lean, fat and bones was obtained by means of the sum of the volumes of the voxels according to its Hounsfield value. Different measurements of fat and muscle thickness and muscle area were obtained from images of the ham and the loin.
Linear, quadratic and allometric regression equations were obtained to predict carcass lean meat percentage and ham composition by means of data obtained from CT images. Linear models including volume plus measurements of muscle and fat thickness and areas, were those that minimized the root mean square error (RMSE) in the prediction of carcass lean meat percentage (1.45 %) and of weight of fat (102.32 g) and bone (25.81 g) of the ham. Ham weight and ham muscle weight were better predicted by quadratic models (RMSE=335.19 g and 245.08 g, respectively) using only tissue volumes as predictors.
Maria Font Furnols, Anna Carabus, Candido Pomar, Albert Brun, Joan Tibau, Marina Gispert. Prédiction du rendement de la carcasse et de la composition du jambon à partir d’images de tomographie à rayons X de porcs vivants durant leur croissance. 46èmes Journées de la Recherche Porcine.
