Minced Meat Classification using Digital Imaging System Coupled with Machine Learning

Meat is widely recognized as a prevalent dietary choice due to its intricate nutritional profile, and its production constitutes a significant component of the global cattle industry. Due to the significant market demand for meat and the pursuit of maximizing profitability, meat products are often adulterated with additional ingredients. A diverse amount of meat can be effectively detected by the comparison model of machine learning. This study provides a comprehensive analysis of the performance of low-cost imaging for classifying minced meat product. In this study, a variety of raw minced meat, including beef, lamb, and pork, was utilized. We investigated the feasibility of low-cost imaging system coupled with machine learning classifier such as Decision Tree, Random Forest, K-Nearest Neighbors (KNN), Support Vector Machine (SVM), Logistic Regression, and AdaBoost Classifier. The findings indicate that the K-nearest neighbors (KNN) model exhibits superior performance, with an overall classification accuracy of 98.2%. The K-Nearest Neighbors (KNN) model exhibits a notable level of precision and accuracy across all classes. Therefore, this study confirms that machine learning algorithm provide robust features for classifying minced meat from the image data.