Corn leaf disease: insightful diagnosis using VGG16 empowered by explainable AI

Area of Science:

• Agricultural Science
• Computer Science
• Artificial Intelligence

Background:

• Corn diseases threaten global food security and economic stability.
• Early and accurate disease detection is vital for high crop yields.
• Deep learning models offer potential for corn disease classification but lack transparency.

Purpose of the Study:

• To develop and evaluate an explainable deep learning model for classifying corn leaf diseases.
• To improve the transparency and reliability of AI models in agriculture.
• To aid farmers in identifying and managing corn crop diseases.

Main Methods:

• Utilized the VGG16 deep learning model for image classification.
• Implemented Layer-wise Relevance Propagation (LRP) to generate heat maps for model interpretability.
• Classified corn leaves into four categories: healthy, blight, gray spot, and common rust.

Main Results:

• The VGG16 model augmented with LRP achieved high accuracy in corn leaf disease classification.
• The model successfully identified critical regions in images, providing human-readable explanations.
• Outperformed previous state-of-the-art models in classification performance and interpretability.

Conclusions:

• The proposed explainable AI approach enhances transparency and reliability in corn disease diagnosis.
• This method empowers farmers with actionable insights for improved crop management.
• Facilitates better decision-making to mitigate crop loss and enhance agricultural productivity.