Agricultural AI: Integrating Federated Learning and CNN in Papaya Leaf Disease Detection

This study paper gives a new way to find and classify diseases in papaya leaves using federated learning with convolutional neural networks (CNN). The research is about finding four levels of lousy health in papaya leaves. It uses information from four clients to do this job well. Using federated learning helps turn local data understanding into a united global view, which is essential for proper disease diagnosis. They used federated averaging to join data findings from each client, creating an overall global model. The model's performance was closely checked using big, small, and weighted averages. Clients' results showed mixed but essential success in classifying diseases. In simple terms, the client named sv_1 had a big average score of 87.84 and smaller averages for each class at 87. It showed consistent performance in various classes equally well. Client SV_2 had slightly better numbers. Its average score was 90.31, and the weighted average shot up to 90.03, showing that it's getting better at finding diseases than others. Client sv_3 did very well in scoring, with a significant average of 94.98, a weighted score of 94.95 and then another measure called micro average, which had the same score as the overall one with nothing above all else like high tech classifications, leading it to be the best among its competitors by showing off such skills you can not ignore or In the end, client SV_4 showed strong results, with an average of 92.89 for all macro things and 92.74 for small micro parts connected to it. Finally, the study gives a detailed plan for correctly finding how serious papaya leaf diseases are using federated learning and CNN. The data and tests show that the model works well and can be trusted, helping to create better solutions in farm disease control that do not depend on one centre.