RICEDX-LIME: Multi-Scale Attention Network with Context-Aware Explainability for Tropical Rice Pathology

DOI

10.2991/978-94-6239-616-6_73How to use a DOI?

Keywords

Rice Leaf Disease Detection; EfficientNet; Deep Learning; Agricultural Diagnostics; Visual Explanation; Model Interpretability; Explainable AI (XAI)

Abstract

In this paper, we present RiceDx-LIME, a deep learning framework for rice leaf disease detection that combines a novel, context-aware LIME-based explainability module (BioLIME) with a hybrid EfficientNet-B4 architecture to improve diagnostic accuracy and model transparency. With an F1-score of 0.91 and a state-of-the-art accuracy of 93.2% across three disease classes, Brown Spot, Leaf Smut, and Bacterial Blight, our system was trained and assessed on a dataset of 3,450 field photos of rice leaves. Compared to current techniques, this indicates a 5.8% increase in accuracy. In addition, the system’s inference time (210 ms) is 28% faster than that of similar models. The core innovation lies in its integrated explainability pipeline, which provides intuitive, colour-coded visual explanations of disease-specific features (e.g., lesions, chlorosis), enabling farmer-friendly diagnosis. By combining excellent performance with actionable interpretability, RiceDx-LIME, which has been validated by agricultural experts in five Indian states, establishes a new benchmark for explainable AI (XAI) in agricultural diagnostics.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - J. Arockia Jackuline Joni
AU  - M. Mary Shanthi Rani
PY  - 2026
DA  - 2026/03/31
TI  - RICEDX-LIME: Multi-Scale Attention Network with Context-Aware Explainability for Tropical Rice Pathology
BT  - Proceedings of the International Conference on Artificial Intelligence and Secure Data Analytics (ICAISDA 2025)
PB  - Atlantis Press
SP  - 994
EP  - 1008
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-616-6_73
DO  - 10.2991/978-94-6239-616-6_73
ID  - Joni2026
ER  -