Early betel leaf disease detection using vision transformer and deep learning algorithms

Betel leaves benefits include their high content of antioxidants, which can help protect against oxidative stress and promote overall health. Additionally, these leaves are known for their potential anti-inflammatory properties, making them valuable in traditional medicine practises. The biggest threat to the security of the food supply is posed by plant diseases, and it is difficult to detect them early enough to prevent potential economic harm. This crop loss not only affects the economy but also poses a threat to food security, as betel leaves are widely used in traditional cuisines and herbal remedies. By analysing large datasets of plant images, deep learning algorithms can quickly identify specific patterns and symptoms associated with various diseases. In this study, we evaluated how well four deep learning models—VGG19, DenseNet201, ResNet152V2, and a Vision Transform model—performed at detecting diseases that affect betel leaves. Both the ResNet152V2 and VIT models attained levels of accuracy, with testing accuracies of 98.42% and 97.83%, respectively. However, the VGG19 model had slightly lower accuracy, with a testing accuracy of 91%. Overall, these deep learning models showed promising results in detecting diseases affecting betel leaves, with the DenseNet201 model performing the best with a testing accuracy of 98.77%. © 2023, The Author(s), under exclusive licence to Bharati Vidyapeeth's Institute of Computer Applications and Management.