Poultry Disease Detection and Treatment Recommendation Using CNNs

This research presents a groundbreaking mobile application to revolutionize poultry disease management. The application uses automated artificial intelligence to accurately diagnose and recommend appropriate treatments for chicken ailments. A robust methodology combines a convolutional neural network (CNN) for visual disease recognition with a rules-based recommendation system to provide tailored action plans. A vast dataset of real-world images depicting both healthy and sick chickens was collected to train the CNN model. Rigorous testing ensured the application’s adherence to ISO 25010 software quality standards, encompassing functionality, reliability, usability, efficiency, maintainability, and portability. An Android prototype demonstrated an impressive 87% accuracy in disease classification. Evaluation trials involving veterinarians and poultry farmers confirmed the effectiveness and relevance of the recommended actions. The fusion of explainable AI with domain expertise distinguishes this innovation from traditional black-box models, significantly reducing the risk of misdiagnosis. Contextual user research highlights the application’s potential to enhance smallholding revenue sustainability. Future directions include model refinement and participatory design to address local specificities and further optimize the application’s impact on poultry health and economic viability.