Reproducibility Resources for Federated Learning-Based Maize Leaf Disease Recognition under Client Heterogeneity and Cross-Domain Variation

Published: 12 October 2026| Version 1 | DOI: 10.17632/zbbhmdxf6m.1
Contributor:

Description

This dataset provides reproducibility resources for evaluating federated learning-based maize leaf disease classification under client heterogeneity and cross-domain variation. It supports experiments using an EfficientNet-B0 model trained centrally and through Federated Averaging (FedAvg) with different client data partitioning strategies, including IID and non-IID distributions. The resources include research code, dataset manifests, federated client partition configurations, evaluation outputs, and analysis materials for assessing model performance on internal and independent external test sets. The experiments use maize leaf disease categories including Common Rust, Gray Leaf Spot, and Northern Leaf Blight, with data sourced from PlantVillage and Plagues_Maiz. These materials are intended to support transparency, reproducibility, and further research into federated learning for agricultural disease recognition. Original dataset images are included only where permitted by the applicable dataset licenses.

Files

Institutions

Categories

Artificial Intelligence, Computer Vision, Federated Learning

Licence