Prediction Of Intramuscular Fat Percentage In Live Swine Using Real-time Ultrasound

Previous study about ultrasound predict Intramuscular fat (IMF)have been progressed in this test.The content of this research included filling of image parameters,and the statistical analysis ofModel.The mean of ultrasound predict IMF percentage have been introduced in the end.Lean pigs (n=25)and obese pigs (n=53)were used to develop model to predict 1MF oflongissimus muscle in live pigs.A minimum of four longitudinal real-time ultrasound images were collected at the last ribs on live animals.Three different methods (untreated,median filter and equalization processing)were used to remove noise in ultrasound images by MATLAB software and 11 candidate image parameters were collected to develop model.After slaughter,a slice fromthe last rib loin interface was used to determine carcass loin intramuscular fat percentage (CIMF).By using CIMF as dependent variable,11 candidate image parameters from the last rib at loinmuscle area in live animal as independent variables,this model conducted linear regression analysistogether with SPSS.The results showed that:1. The gray value of b-ultrasound image has good quality with high precision and little measurement error,while the repetition was 0.954.One ultrasound image can be used to predictIMF in live swine.2. The best effect of regression equation could get in lean pigs after removing noise by equalization processing,The optimum regression equation wasy=-1.125+0.012×P₄The R,R² and adjusted R² was 0.675,0.456 and 0.432 respectively,(y was the predictivevalue of model,P₄ represent gray value in x₂ to x₂₊₅ on special area of b-ultrasound image,(?)₂ was average gray value)3. The best effect of regression equation could get in lean pigs after removing noise by medianfilter.The optimum regression equation wasy=6.405-0.012×P₄The R,R² and adjusted R² was 0.392,0.154 and 0.137 respectively...