Identifying Pork Raw-Meat Based on Color and Texture Extraction Using Support Vector Machine

Meat is one of the foods that are rich in animal protein. The demand for meat in Indonesia continues to increase each year in line with the increase in population. Many traders are deliberately mixing beef with pork to get big profits. The utilization of technology can be applied to solve this problem by using image processing. This research was made to see the performance of Support Vector Machine (SVM) in identifying the difference between beef with pork based on color and texture characteristics with simulation in Matlab 2015a. Color feature extraction uses RGB-to-HSV transformation, which results in hue, saturation, and value. Texture feature extraction using Local Binary Pattern (LBP) transformation. The results of this LBP will be processed on the extraction of first-order statistical features, resulting in mean, entropy, variance, kurtosis, and skewness. And from 8 parameters of color and texture extraction value, then used as input for classification using SVM. The highest accuracy achieved in this research was 90%, using a total of 30 data images of beef, pork, and wild boar.