Pioneering Crop Health in Agriculture: Federated Learning CNNs in Papaya Leaf Disease Detection

This research paper uses Federated Learning (FL) with Convolutional Neural Networks (CNNs) to classify papaya leaves into five diseases. This research is carried out collaboratively with the help of The key objective here is to utilize the decentralized nature of FL to improve the accuracy and efficiency of disease classification while ensuring data privacy. We used federated averaging techniques to combine all the data sets and translate the local information into a robust global model. It is commonly used for averaging different averages. The results showed that every client had high accuracy and recall. The clientva_1 achieved excellent results with precision and recall rates of 80.63% and 80.69%, respectively, with an F1-score of 80.55% and an accuracy of 92%. Client va_2 achieved excellent results with precision and recall rates of 83.03% and 83.05%, an F1-score of 83.01%, and an accuracy of 93%. The precision and recall of the client va_3 were 78.29% and 78.24%, respectively, and its F1-score was 78.18%. Its accuracy was reported to be 91%. From my analysis, the client va_4 displayed better results in precision and recall rates with values of 84.57% and 84.61%, a high F1-score of 84.53%, and an accuracy of 94%, indicating excellent performance.In conclusion, client va_5 attained precision and recall rates of about 83.09% and 83.35% and an F1-score of approximately 83.16%, with an accuracy of almost 93%. In the case of the group analysis, for macro averages across the participants, the results ranged between 78.24% and 84.57%; for weighted averages, the values fluctuated between 78.17% and 84.53%; and for micro standards, the scores varied from 78.08% to 84.49%. The consistent and balanced performance across different averaging methods shown in the results shows the model's effectiveness in disease classification.