MCWDPC: Multi-Class Wheat Diseases Prediction and Classification Using Deep Learning Hybrid Model

This study presents the Hybrid Feature Fusion and Lightweight CNN (HFF-LCNN) framework for wheat disease classification, achieving an accuracy of 97%, precision of 97%, recall of 98%, and an F1-score of 97%. The model leverages deep learning with hybrid feature fusion to capture diverse disease attributes, while an attention mechanism focuses on diseasespecific features. A modified ResNet152 backbone, combined with Explainable AI (XAI) techniques, ensures high accuracy and interpretability. The hybrid segmentation approach effectively isolates disease regions, enabling robust agricultural diagnostics even in complex environments. The lightweight design of HFFLCNN supports real-time classification, making it suitable for resource-constrained settings. By outperforming existing models by 2-5% in accuracy, this framework offers a transformative tool for precision farming, empowering farmers with reliable and actionable insights to enhance crop management and minimize yield losses. This research sets a benchmark for real-time, efficient wheat disease detection.