Fat Estimation in Beef Ultrasound Images Using Texture and Adaptive Logic Networks

Abstract

Adaptive Logic Networks (ALNs) and co-occurrence image texture are used for both prediction and classification of intramuscular fat in beef from ultrasonic images of both live beef animals and slaughtered carcasses. Prediction results showed that ALNs had a mean error between 0.83 and 0.94% fat, and classification accuracy, according to major grade divisions, ranged from 73 to 79%. The ALNs are a viable alternative to statistical techniques.