IMAGE ANALYSIS OF BEEF
Figure 1. Images of a steak before analysis (top left), with the fat highlighted in black (top right), with the lean highlighted in black (bottom left), and with the biggest lean piece (EYEPIECE) highlighted in black (bottom right). Note that EYEPIECE is not necessarily limited to the longissimus.
wholesale cut (round, loin, rib, chuck, flank, and brisket/plate/foreshank) was individually dissected and the following components were weighed: 1) boneless, totally trimmed retail cuts, 2) fat trim, 3) lean trim, and 4) bone. Weights of lean and fat trim were adjusted to a constant 20% fat lean trim basis. Weights of boneless, totally trimmed retail cuts and 20% fat lean trim were summed to give retail product weight ( RPWT). Retail product yield ( RPYD) was expressed as a percentage of the sum of the parts (i.e., RPYD = 100 RPWT/[RPWT + fat trim weight + bone weight]) rather than as a percentage of hot carcass
weight ( HCW) to overcome potential introduction of error that was due to shrink and(or) cutting loss.
Statistical Analysis. Carcasses were blocked by observed RPYD, and one-half of the carcasses were used to develop regression equations and one-half of the carcasses were used to validate the regression equations (Neter et al., 1989). Regression equations were developed for RPYD and SLA. Retail product weight was predicted by multiplying the best estimate of RPYD times HCW. All of the image analysis traits described above and HCW were included as potential independent variables. Regression equations were
a0 = black; 255 = bright red, green, or blue.
Table 1. Histogram ranges for measuring lean area, fat area, and total steak area
Brightnessa Green Minimum Maximum