The poultry industry is experiencing substantial growth globally, leading to increased demand for high-quality poultry products. However, this growth has also been accompanied by a rise in food fraud incidents, which has raised concerns about the integrity and safety of poultry products within the supply chain. As a result, there is a pressing need for advanced solutions that can accurately assess the quality of poultry production and detect any potential diseases or abnormalities in meat. This study proposes the use of deep learning algorithms for the quality assessment of poultry and livestock meat, with a focus on disease detection. We employ three advanced machine learning models—ResNet-18, EfficientNetB3, and MobileNet—trained on a publicly available dataset from Kaggle, which contains images of poultry diseases. These models are evaluated based on their performance in classifying and detecting various poultry diseases, with key metrics such as accuracy, precision, and recall being used for comparison. The results of this study demonstrate that deep learning models can effectively detect poultry diseases with high accuracy, offering a promising solution for improving food safety and reducing fraud within the poultry supply chain. By integrating these models into the production and quality assurance processes, the industry can ensure more reliable and safe poultry products for consumers.

Keywords: Poultry Industry, Quality Assessment, Meat Detection, Food Fraud, Machine Learning, Deep Learning, ResNet-18, EfficientNetB3, MobileNet, Disease Detection, Food Safety, Artificial Intelligence.