Strawberries at the Nexus of Tech: Federated Learning and CNN for Disease Severity Classification

The timely and precise identification of plant diseases is still crucial in precision agriculture to guarantee maximum crop health and yields. Using a novel combination of Federated Learning and Convolutional Neural Networks (CNN), this study sets out on a cutting-edge quest to assess the subtle severity levels of strawberry leaf illnesses. The research explores four well-defined severity levels using datasets from six clients, each capturing various aspects of the disease's course. While CNN design is optimized to identify even the most delicate morphological signs of the illness, the decentralized nature of Federated Learning guarantees data privacy. The resulting Federated Averaging CNN model displays excellent performance metrics. Macro averages range from 78.64% for the hardest client datasets to 92.04% for the pickiest ones. Weighted averages range from 78.89% to 92.02%, whereas micro standards, which vary from 78.90% to 92.01%, demonstrate a comparable level of performance. The model's clever ability to categorize the severity of strawberry leaf ailments is shown by these measures alone and together. This study represents a breakthrough in plant disease identification by bringing a practical, scalable, and effective solution to strawberry growers worldwide while simultaneously breaking new ground in machine learning and agronomy.