Tomato Leaf Image Dataset for Disease Analysis in Real-World Environment

The dataset comprises a total of 1,028 images of tomato leaves collected in Bangladesh, with 486 images depicting healthy leaves and 546 images showing diseased leaves. Each image has been carefully annotated to indicate specific regions as either healthy or diseased, given that each image includes a complex background. The annotations are provided as text files, where each image file has a corresponding text file with the same name to facilitate model implementation. Healthy images are labelled with "H," while diseased images are labeled with "D." For structured model development, the dataset is divided into two folders: a validation folder and a training folder. The validation folder contains 160 images, equally split between 80 healthy and 80 diseased leaves, while the training folder includes the remaining 868 images. This dataset is particularly valuable for training and validating deep learning and machine learning algorithms aimed at detecting diseases in tomato leaves. It offers researchers and learners a robust resource for analysing and improving the health management of tomato plants through the development of advanced computational models.