ResPoultry: An Enhanced ResNet50 Model for Multiclass Classification of Poultry Diseases

Timely and accurate diagnoses of ailments affecting poultry health are crucial to achieving food security in addition to promoting environmentally friendly agricultural development. This research uses ResNet50, a latest-generation artificial neural network model that is considered the vanguard of technology in this field, to skillfully detect all kinds of poultry diseases. Extensive testing of combinations of different learning rates and the Adam optimizer as well as SGD has been carried out methodically. These results provide further proof that the model is working fine, especially if you use Adam with an initial learning rate of 0.01 to train it. In a particularly noteworthy performance, the model scored an accuracy of 97.03 % during testing and 98.75 % during training (see table). Furthermore, the lowest training loss is found to be 0.024 with a testing loss of 0.16 at learning rate = 0.01 original size The tests appear promising so far for two reasons; one results from the use of augment in baseball and another comes about as an accident due to increased motivation on their part though I should not think this way because things are different. Such a proposed model promises additional improvement to permit the identification of even more poultry diseases. In addition, considering integration with modern-day farm equipment will allow for the strengthening of real-time disease monitoring functions. This would support timely action to address poultry-related disease problems, and thus strengthen the resilience of the entire industry. In addition to helping strengthen food security, these concerted efforts also help protect the livelihoods of poultry farmers--generally small-scale family growers. As the work continues, it provides a basis for ongoing improvement in poultry disease prevention and control essential element of lasting strength and stability to poultry farms.