Classification of Sweet Potato Leaf Variants using Transfer Learning

The sweet potato (Ipomoea batatas) is a significant crop with a wide range of leaf color, texture, and shape. The categorisation of these variants must be exact in terms of agricultural methods and crop management. This project attempts to create a reliable classification system for sweet potato leaf varieties using deep learning methods. The study examines the categorization performance of two well-known deep learning models, VGG16 and Inception v3. The models are trained using a dataset that consists of tagged photos of numerous sweet potato leaf varieties. By using weights that were previously trained on ImageNet, transfer learning is put into practice. The performance of the models is assessed using accuracy and the F1 score. According to experimental findings, the VGG16 and Inception v3 models both exhibit high accuracy rates, with VGG16 achieving 95.57 percent accuracy and Inception v3 achieving 96.97 percent accuracy. These findings show that deep learning has the potential to accurately classify variations in sweet potato leaves. By creating an effective method for automating the classification of sweet potato leaf variants, this study makes a contribution to the field of agricultural research that can be used to enhance crop management and breeding initiatives.