Prediction of IMF Percentage of Live Cattle by Using Ultrasound Technologies with High Accuracies

The purpose of this study is to produce algorithms that are able to predict the intramuscular fat (IMF) percentage of live cattle. Two algorithms based on linear regression analysis and neural network models are developed. These two algorithms extract feature information from live cattle ultrasound images and calculate the predicted IMF percentage values. The results show that these algorithms perform better than the previous studies in the same field. A brief description of the data acquisition process, the ROI extraction, the mathematics of the feature selection methods, statistical analysis on P-value and correlation, and the outputs from Matlab programs is presented.