A Hybrid Deep Learning Approach for Accurate and Transparent Maize Plant Disease Classification

Within the domain of agricultural progress, accurately identifying instances of plant diseases emerges as a pivotal hurdle. We introduce a fresh approach to classify diseases in maize plant leaves by combining a hybrid model that merges Convolutional Neural Networks (CNNs) for visual recognition with Multi-Layer Perceptrons (MLPs) for transparency. This amalgamation aims to bridge the gap between precision and comprehensibility. Our study begins by scrutinizing the constraints of prevailing CNN and MLP-based models in the context of disease classification. Subsequently, we present our hybrid architecture, elucidating its benefits and addressing inherent challenges. Through rigorous experimentation, we showcase the model's remarkable performance, achieving training, validation, and test accuracies of 98.98%, 96.19%, and 95.76%, respectively. The significance of our proposed model lies in its potential to transform disease management in agriculture. The fusion of accuracy and interpretability not only equips farmers with actionable insights but also establishes a model for ethical and transparent AI implementation. This research propels the field forward by providing an inventive hybrid solution that excels not only in accuracy but also harmonizes with practical agricultural scenarios, thus paving the path for a more sustainable and enlightened future.