![[2602.12484] A Lightweight and Explainable DenseNet-121 Framework for Grape Leaf Disease Classification](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.12484] A Lightweight and Explainable DenseNet-121 Framework for Grape Leaf Disease Classification
Summary
This article presents a novel DenseNet-121 framework for classifying grape leaf diseases, achieving high accuracy and interpretability while being computationally efficient.
Why It Matters
Grape diseases significantly impact agricultural productivity and quality. This research offers a lightweight and explainable solution for early disease detection, which is crucial for sustainable vineyard management. The findings could enhance the adoption of AI in agriculture, making it more accessible and effective.
Key Takeaways
- The proposed DenseNet-121 framework achieves 99.27% accuracy in grape leaf disease classification.
- Utilizes Grad-CAM for interpretability, enhancing trust in AI decisions.
- Optimized for real-time deployment with reduced computational costs.
Computer Science > Computer Vision and Pattern Recognition arXiv:2602.12484 (cs) [Submitted on 12 Feb 2026] Title:A Lightweight and Explainable DenseNet-121 Framework for Grape Leaf Disease Classification Authors:Md. Ehsanul Haque, Md.Saymon Hosen Polash, Rakib Hasan Ovi, Aminul Kader Bulbul, Md Kamrul Siam, Tamim Hasan Saykat View a PDF of the paper titled A Lightweight and Explainable DenseNet-121 Framework for Grape Leaf Disease Classification, by Md. Ehsanul Haque and 5 other authors View PDF Abstract:Grapes are among the most economically and culturally significant fruits on a global scale, and table grapes and wine are produced in significant quantities in Europe and Asia. The production and quality of grapes are significantly impacted by grape diseases such as Bacterial Rot, Downy Mildew, and Powdery Mildew. Consequently, the sustainable management of a vineyard necessitates the early and precise identification of these diseases. Current automated methods, particularly those that are based on the YOLO framework, are often computationally costly and lack interpretability that makes them unsuitable for real-world scenarios. This study proposes grape leaf disease classification using Optimized DenseNet 121. Domain-specific preprocessing and extensive connectivity reveal disease-relevant characteristics, including veins, edges, and lesions. An extensive comparison with baseline CNN models, including ResNet18, VGG16, AlexNet, and SqueezeNet, demonstrates that the propose...
