Enhancing sugarcane leaf disease classification using vision transformers over CNNs

Abstract Sugarcane is a globally significant crop facing threats from leaf diseases that impact its productivity. Traditional detection methods are often inefficient and time-consuming. This study explores the use of Vision Transformers (ViT) for classifying sugarcane leaf diseases and compares thei...

Full description

Saved in:
Bibliographic Details
Main Authors: Saritha Miryala, Krupa Rasane
Format: Article
Language:English
Published: Springer 2025-06-01
Series:Discover Artificial Intelligence
Subjects:
Online Access:https://doi.org/10.1007/s44163-025-00340-7
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Abstract Sugarcane is a globally significant crop facing threats from leaf diseases that impact its productivity. Traditional detection methods are often inefficient and time-consuming. This study explores the use of Vision Transformers (ViT) for classifying sugarcane leaf diseases and compares their performance with traditional CNNs. A dataset of 19,926 images across six classes was used to fine-tune both ViT and CNN models. The optimized ViT model achieved a test accuracy of 96.53%, outperforming the CNN models (ResNet50 and VGG16) with accuracies of 91.92% and 92.30%, respectively. These findings demonstrate the superior performance of ViTs over CNNs in early disease detection for sustainable crop management. Future work will focus on expanding the dataset and optimizing model parameters for further improvements in disease classification accuracy.
ISSN:2731-0809