article · IEEE Access
Plant diseases pose severe risks to agricultural yields, product quality, and the broader economy. While deep learning offers strong capabilities for automated disease recognition, the opaque nature of these models makes their predictions difficult to trust and validate. To resolve this issue, an explainable artificial intelligence system was developed to classify plant conditions accurately while offering clear rationales for its outputs. The system detects 38 separate plant diseases, achieving 99.69 percent accuracy, 98.27 percent precision, and 98.26 percent recall. Visual explanations for the predictions are generated using the local interpretable model-agnostic explanations framework, ensuring that the machine decisions align with established explanatory practices and domain knowledge. This transparent approach aims to support dependable decision-making in agricultural disease monitoring.
Crop diseases directly reduce agricultural productivity and undermine global food security. Although deep learning can spot these diseases swiftly, agricultural managers often hesitate to act on predictions they cannot understand. Coupling high-precision diagnostics with visual explanations allows users to see exactly why an ailment was diagnosed, facilitating faster and more confident interventions in the field.
This applied research is tested on a diagnostic dataset of 38 plant diseases and appears to be in an early to intermediate testing stage. It could eventually enable transparent disease-monitoring tools for agronomists, crop advisers, and digital farming platforms. Further development would be required to integrate the explanatory model into deployable farm software or handheld diagnostic devices.
AI-generated from the published abstract. Always read the original work before citing.
Plant diseases can have profound effects on the economy, impacting both local and global scales. These diseases can lead to substantial losses in agricultural productivity, affecting crop yields and quality. In this context, deep learning algorithms are widely acknowledged as effective solutions. However, the use of these black-box approaches raises concerns about trust in interpreting and validating the decisions generated by the models. This study proposes an explainable artificial intelligence (XAI) based plant disease classification system to classify and identify distinct ailments with improved accuracy. The system correctly identifies 38 different plant diseases with accuracy, precision, and recall as 99.69%, 98.27%, and 98.26%, respectively. These predictions are subjected to additional analysis employing the local interpretable model-agnostic explanations (LIME) framework to produce visual explanations aligning with prior beliefs and adhering to established best practices in explanations. This system will serve as a promising avenue for revolutionizing disease detection, fostering informed decision-making, and ultimately contributing to global food security.
This page summarises published work. The authoritative version sits with the publisher.
DOI: 10.1109/access.2024.3428553
Is something wrong with this record? Report it or request removal.
Discussion
Have you built on this work, tried to replicate it, or seen it applied in practice? Share what you know. Verified researchers and MARATTO™ domain experts can open a discussion, and any member can reply. Contributions are reviewed before they appear.
No discussion yet. Open the first thread.
New to MARATTO™? Create a free account.
