,
Research Scholar & Assistant Professor, GITAM School of Computer Science & Engineering, GITAM (Deemed to be University), Doddaballapur, Bengaluru, Karnataka, India
Professor, GITAM School of Computer Science & Engineering, GITAM (Deemed to be University), Doddaballapur, Bengaluru, Karnataka, India
The study is based on the idea of an innovative method of automated classification of mulberry root Knot diseases on the basis of the leaf symptoms through the EfficientMulberryNet-B0-GC model, which is a combination of EfficientNet-B0 and Grad-Cam visualization. The model aims to complement the lack of efficient and transparent disease detection in agricultural practices and offer a strong solution for real-time classification of mulberry root knot disease according to leaf symptoms. The method proposed uses a dataset of mulberry leaves that consists of high-resolution images of two classes, i.e., Healthy, Root knot disease. The EfficientMulberryNet-B0-GC model, which is pre-trained on ImageNet, is then fine-tuned to classify diseases with data augmentation techniques used to enhance the generalization of the model. It makes the model's decision-making process transparent and more credible by using the technique of Grad-CAM to visualize and interpret the model's decision-making process. The results indicate that the proposed model outperforms existing models as it attains 94% accuracy, 95% precision, 95% recall, 95% F1 score, and 99% AUC. The data set is of 2500 images, diversified to be strong and extensive. The framework provides a scalable, practical approach to disease monitoring to provide improved disease accuracy, interpretability, and reliability. The study helps to develop AI-based tools in the agricultural sector, especially the detection and management of diseases in the mulberry.
This is an open access article distributed under the Creative Commons Attribution Non-Commercial License (CC BY-NC) License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The statements, opinions and data contained in the journal are solely those of the individual authors and contributors and not of the publisher and the editor(s). We stay neutral with regard to jurisdictional claims in published maps and institutional affiliations.
