Disease Identification in Potato Leaves using Swin Transformer

Abstract

One of Indonesia's mainstay agricultural products is the potato. Disease prevention is essential for maintaining stable potato production. One technique for detecting disease in potatoes is to determine whether potato leaves are diseased (early blight or late blight) or healthy. Deep Learning models have been widely developed and used to classify disease recognition in potato leaves in the industrial era 4.0. Swin Transformer is a deep learning model based on transformers that is more efficient and accurate at solving classification problems. The Swin Transformer, a transformer based deep learning approach, is used in this study to identify diseases of the potato leaf. Moreover, several metrics including Precision, Recall, Accuracy, and F1 score, are used to assess the experimental results of the model we use. In terms of accuracy, the value obtained when training with this model is 97.70%. These findings indicate that using the Swin Transformer model to identify potato leaf diseases could be a new trend in agricultural research.