Predicting of Beef Carcass Marbling Using the Digital Information by Ultrasonic Image Analysis

Abstract

Ultrasonic technique is widely used in predicting marbling score in live beef cattle. The judging techniques were, however, done subjectively on hardcopy or videotape recorder (VTR) image. In this study, the digital information obtained from the ultrasonic scanogram was used to estimate the beef marbling score (BMS) and to clarify the possibilities for predicting the BMS objectively by image analysis. Four hundred and ninety nine fattening cattle were ultrasonically scanned (HS-2000), and the digital information for predicting the BMS were interpreted on the computer using a commercial image analysis software. It was shown that some digital informations on M. longissimus thoracic and M. semispianlis dorsi would decrease correspondingly with the increase of estimating the BMS on the carcass. The result of multiple regression analysis for estimating the BMS as the dependent variable and digital informations obtained by image analysis as the independent variable showed that 9 independent variables were the basis for predicting the BMS and the contribution ratio of the equation was 0.603. 71.5% ±1 was in agreement for estimating the BMS.