Toward Reliable Tea Leaf Disease Diagnosis Using Deep Learning Model: Enhancing Robustness With Explainable AI and Adversarial Training

2026-02-13 · Source: cs.CV updates on arXiv.org · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Precision Agriculture & Smart Farming · Depth: Intermediate, quick

Summary

A study developed an automated deep learning model for tea leaf disease classification using the teaLeafBD dataset, which comprises 5,278 high-resolution images across seven categories, including six disease types and healthy leaves. The research employed DenseNet201 and EfficientNetB3 models within a pipeline that incorporated data preprocessing, splitting, augmentation, adversarial training, and Explainable AI (XAI) strategies like Grad-CAM visualization. Adversarial training was specifically applied to enhance model robustness against noisy or disturbed inputs. Experimental results showed EfficientNetB3 achieved a 93% classification accuracy, outperforming DenseNet201, which reached 91%. This approach aims to provide an efficient and accurate method for detecting tea leaf diseases, supporting advanced agricultural management.

Key takeaway

For agricultural managers or AI scientists developing plant disease detection systems, this research demonstrates that integrating adversarial training and Explainable AI with models like EfficientNetB3 can significantly boost diagnostic accuracy and robustness. Consider adopting these techniques to improve the reliability of your automated systems, especially when dealing with potentially noisy real-world image data, and to gain insights into model predictions.

Key insights

Deep learning with adversarial training and XAI enhances tea leaf disease detection robustness and interpretability.

Principles

• Adversarial training improves model resilience to noisy inputs.
• XAI methods like Grad-CAM enhance model interpretability.

Method

The pipeline involves data preprocessing, splitting, augmentation, adversarial training, model training, evaluation, and Grad-CAM visualization for explainability.

In practice

• Use EfficientNetB3 for high accuracy in image classification.
• Apply adversarial training to improve model robustness.
• Employ Grad-CAM for visualizing model decision regions.

Topics

• Tea Leaf Disease Diagnosis
• Deep Learning
• Explainable AI
• Adversarial Training
• EfficientNetB3

Best for: Computer Vision Engineer, AI Scientist, Research Scientist, AI Engineer, Machine Learning Engineer, Data Scientist

Related on AIssential

• An Ensemble Deep Learning Approach for Reliable and Scalable Lemon Leaf Disease Classification — An ensemble deep learning approach, integrating InceptionV3 and MobileNetV2 with adversarial training and Grad-CAM, reliably classifies lemon leaf diseases with 99.27% accuracy.
• DeepGreen: a real-time deep learning system for smart agriculture monitoring — A novel Conv-7 DCNN with modified ParNet attention achieves high accuracy and real-time performance for plant disease detection.
• Comparison of Deep Learning Frameworks For Rice Disease Mapping From UAV Multispectral Imaging — Lightweight CNNs excel in UAV-based rice disease mapping, with U-Net++ EfficientNet-B3 achieving 97.62% mIoU.
• BrainFusionNet: a deep learning and XAI model to understand local, global, and sequential features of MRI images for improved brain tumour detection — BrainFusionNet combines CNNs, ViT, and GRUs with XAI for highly accurate and interpretable brain tumor detection from MRI.
• Batch-Invariant Spectral Intelligence for Robust and Explainable Insect Authentication — BISN uses learnable preprocessing and adversarial training to achieve robust, explainable insect authentication despite batch variations.
• Explainable AI-Driven Cyber Risk Analytics and Model Reliability Assessment for Intelligent Governance of U.S. Critical Infrastructure: An XGBoost and SHAP-Based Intrusion Detection Framework — Integrating XAI with ML-based intrusion detection enhances transparency and trust for critical infrastructure cyber governance.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by cs.CV updates on arXiv.org.