An Efficient CNN-based Model for Meat Quality Assessment: The Role of Artificial Intelligence Towards Sustainable Development

Abstract

The meat industry's pursuit of consistent and accurate quality assessment is essential to meet consumer demands and ensure food safety. Meat freshness is an important consideration when evaluating product quality and safety and has a significant impact on a consumer's purchasing decision. Traditional techniques that depend on visual inspection by humans are arbitrary and prone to discrepancies. In this paper, we propose a new deep learning architecture model based on convolutional neural networks (CNNs) for meat freshness classification. The model was trained on a meat images dataset to classify the red- meat images into two classes named "fresh" or "spoiled”. The proposed model performance is compared against a set of deep transfer learning models. The obtained results show that the proposed model outperformed other models and achieved the best accuracy of 100% and 100% precision and recall in meat images quality classifications.