Fresh Meat Classification using Image Processing Technique

Meat is considered one of the most nutritious animal food products as it contains carbohydrates, proteins, lipids, vitamins, and minerals. The level of freshness in meat plays a significant role in determining its quality for consumption. It is therefore crucial to maintain meat quality to ensure that consumers receive high-grade meat. Traditionally, meat quality assessment has been done visually by comparing actual meat with reference photographs of each meat class. However, this subjective and time-consuming process has its limitations. To overcome these limitations, an automated image processing-based system is required to detect and assess meat quality. In this project, a convolutional neural network (CNN) approach will be implemented to detect the freshness of meat. The meat samples used in the project will be categorized into three types: fresh, half-spoiled, and spoiled. The objective of this project is to develop an image-processing method that analyzes digital pictures to determine the quality of meat. The system's implementation aims to achieve consistent and objective measurements; as different human examiners may yield different results. This will be accomplished by developing an image classification system that utilizes the CNN algorithm along with image processing techniques. By utilizing the developed model, it can be concluded that the method is effectively applied in the classification of meat freshness with an accuracy rate of 80%.