Explainable AI-Driven Model for Paddy Disease Detection and Classification

Paddy diseases significantly impact the economy on both local and global levels by reducing crop yield and quality. To address this, AI techniques have appeared as powerful tools for disease detection. However, their black-box nature often limits trust in their decision-making processes. This study introduces a paddy disease classification technique based on explainable artificial intelligence (XAI), designed to improve both accuracy and interpretability. The proposed model accurately classifies 13 various paddy diseases, achieving 99.1 % accuracy. To ensure transparency, the system incorporates the LIME framework, which provides visual explanations that align with domain knowledge and accepted interpretability standards. This approach holds great potential for transforming disease diagnosis, supporting data-driven decisions, and advancing global food security.