A Comparative Analysis of Potato Leaf Disease Classification with Big Transfer (BiT) and Vision Transformer (ViT) Models

Potato plants are vulnerable to various leaf diseases that can significantly impact crop yield and quality. Timely detection and accurate diagnosis of these diseases are crucial for effective management and intervention. However, farmers' manual monitoring practices are laborious and prone to errors. Therefore, there is a need for automated approaches to efficiently and accurately detect and classify leaf diseases. In recent years, machine learning techniques and convolutional neural network (CNN) models have shown great potential in automating disease detection in agriculture. However, as technology advances, new approaches such as Big Transfer (BiT) and Vision Transformer (ViT) have emerged as state-of-the-art technologies that hold promise for disease classification tasks. Nevertheless, their application in potato leaf disease classification remains limited. To address this gap, this research aims to classify potato leaf diseases using BiT and ViT models and evaluate their performance based on precision, recall, and f1-score. The study compares the results of both types of models to determine their effectiveness in classifying potato leaf diseases. The findings reveal that BiT outperforms ViT regarding classification performance on both the training (100% precision, 100% recall, 100% f1-score) and test data (97.329 % precision, 97.624 % recall, 97.469 % f1-score).