Using Image Analysis to Estimate Boxed Beef Value

Eighty-four grain-finished steers were slaughtered, evaluated for composition characteristics via image analysis, and fabricated into boxed beef wholesale cuts to determine which measured composition indicators present at the crosssection between the 12th and 13th ribs would most accurately predict boxed beef value. Prediction equations were developed to estimate carcass side value, value per 45.35 kg, primal round value, primal loin value, primal rib value, primal chuck value, and collective primal brisketplate- flank-shank value. Composition indicators chosen to predict boxed beef value included side hot carcass weight; marbling score; LM length, width, and area; subcutaneous fat thickness and area; iliocostalis muscle area; ratio of subcutaneous fat thickness to side hot carcass weight; ratio of LM area to side hot carcass weight; ratio of LM area to subcutaneous fat area; ratio of LM width to subcutaneous fat thickness; and ratio of LM area to subcutaneous fat thickness. Variables frequently correlated with boxed beef value included side hot carcass weight; LM length, width, and area; and subcutaneous fat area. Adjusted R2 values for predicting boxed beef value ranged from 0.60, for estimating beef carcass value per 45.35 kg, to 0.94, for estimating beef carcass side value. These results indicate that boxed beef value can be predicted via existing technology used to augment USDA beef carcass grading.