Tools
Tools
-, Varun Chand H., Mary Kurian, Simy, Vijayan, Asha, Sabharwal, Seema, Kumar, Devipriya S. and Vamadevan, Arundev (2026) The effects of Class Imbalance Handling on Sugarcane Leaf Disease Classification using ResNet-50. In: Proceedings of the 2026 International Conference on Emerging Trends in Information, Communication and Systems, ICETICS 2026. IEEE, Bhopal, India, pp. 1-6. ISBN 979-8-3315-6198-7
Full text not available from this repository.Abstract
Sugarcane is a significant commercial crop and identifying leaf diseases at an early stage is important to minimize losses in yield and sustainability in agriculture. Despite promising performance of deep learning-based methods in automated classification of plant diseases, most of the current research regarding the subject matter does not sufficiently take into account the problem of disparities in the number of classes in real world agricultural data, which results in biased learning and low accuracy in identifying minority disease classes. The paper compares the effect of the class imbalance management carried out in the classification of sugarcane leaf disease using a ResNet-50 convolutional neural network between the traditional training without taking any measures in balancing the class and class-weighted training to overcome the unbalanced distribution of classes. The use of class weights for any dataset of the sugarcane leaf diseases is demonstrated to yield a consistent overall classification performance in terms of accuracy, macro-averaged precision, recall, F1-score and Cohen kappa, through extensive experimental trials. Additionally, analysis by class, confusion matrix analysis and ROC curve analysis confirm high performance enhancements for underrepresented disease groups and more balanced prediction patterns. The results highlight the importance of class-weighted learning in developing reliable deep learning-based agricultural disease classification systems.
| Item Type: | Book Section |
|---|---|
| Uncontrolled Keywords: | class imbalance; class weights; deep learning; plant disease classification; ResNet-50; sugarcane leaf disease |
| Subjects: | Q Science > QK Botany S Agriculture > SB Plant culture Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning |
| Divisions: | School of Computing > Staff Research and Publications |
| Depositing User: | Tamara Malone |
| Date Deposited: | 22 Sep 2026 15:47 |
| Last Modified: | 22 Sep 2026 15:47 |
| URI: | https://norma.ncirl.ie/id/eprint/9939 |
Actions (login required)
 |
View Item |