Enhancing English Oak Disease Detection and Classification Using CNN-RNN Hybrid Model: A Comprehensive Five-Class Approach

This study introduces a novel hybrid Convolutional Neural Network (CNN)-Recurrent Neural Network (RNN) model tailored for the precise detection and classification of diseases affecting the esteemed English Oak species. Development of the model necessitated a meticulously curated dataset comprising 10,000 high-resolution images sourced globally, representing diverse manifestations of disease. Remarkably, the model achieved an impressive overall accuracy of 96.6%, demonstrating its adeptness in accurately delineating severity classes. Evaluation metrics, including precision, recall, and F1-score, underscored the reliability and efficacy of the model, particularly its exceptional capability in identifying severe disease manifestations. However, nuanced observations from the confusion matrix revealed challenges in distinguishing between immediately bad and mild cases below a 0.5 threshold during training. Comparative analysis with existing methods showcased substantial improvements in forest pathology enabled by the proposed model. This tool holds significant potential for early disease identification and precise severity classification, thereby aiding informed forest management strategies. The study highlights the model's merit in advancing subtle disease severity assessment and identify the importance of diverse and updated datasets for its ubiquitous utility and general applicability.