Beef Quality Assessment using InceptionNet

Red Meat is a major supplier of essential proteins and nutrients. It is widely consumed by people around the globe and has been consumed by humans for generations. Subjective quality analysis of red meat is a difficult process. Encountering deep-learning models for the quality evaluation purpose can help quality assessment in a big way. In this research, InceptionNet has been used to classify beef images into two classes – Fresh and Rotten. A dataset of 1000 images has been used and the quality evaluation on this dataset has been carried out using three different validation split ratios. It was found that out of all test split ratios, 80:20 split has achieved the highest accuracy of 96.00% with an epoch number of 20. Future research could explore optimization, larger datasets, and practical applications to improve the prediction of beef quality across many categories. Advanced neural network architectures show promise in the assessment of meat quality.