Agriculture Advances: Federated Learning CNN's for Combatting Banana Leaf Diseases

The study combines federated learning with convolutional neural networks (CNNs) to solve a ubiquitous issue-the banana leaf diseases that dramatically affect global agricultural output. The study focuses on five customary nasal illnesses that consumers suffer, which are modelable and provide analytic information while protecting their data from an attacker. In a federated learning setting, each model is trained using data specific to the client. These parameters are then fine-tuned to create a robust global model. Each client's on-site data (kx_1 to kx_5) is specifically examined to determine accuracy, recall, and Fl-score assessment. Then, all the results are combined to reach the aggregate using federated averaging methods. Micro, macro, and weighted average precisions are determined to put together four performance measures that consider class imbalance and classify whole results more comprehensively. Our results show that the diseases are being classified with high accuracy, which is a significant factor in diagnosis. For keyword x _1, the macro average precision achieved 94.84%, which increased to 96.32% for keyword x_5; the weighted average accuracy went up from 94.3% to 96.32%. Micro-average precision reached 96.32%, compared to 94.82 % in the previous one. It is highest in the case of kx_5. This demonstrates the model's ability to get the correct knowledge from multiple data resources. The findings show federated learning is a solution that fits well to agricultural situations and is capable enough to bring a breakthrough for plant disease detection by being a scalable, efficient, and privacy-aware solution for a global agricultural problem.