Taiwan Tomato Leaves Dataset

Description:

The Taiwan Tomato Leaves Dataset is an extensive and diverse collection tailored for research in plant pathology. With a particular focus on tomato leaf diseases. This dataset comprises 622 meticulously curated images, categorized into six distinct groups: five representing different tomato leaf diseases and one category denoting healthy leaves. These images provide a comprehensive resource for machine learning and computer vision applications. Especially in agricultural disease detection.

The dataset includes a variety of visual scenarios. Such as single leaf images, multiple leaf images, and leaves against both simple and complex backgrounds. The dataset's diversity in composition ensures a robust foundation for developing and testing disease detection models. Furthermore, the images in this dataset vary in their original dimensions but have been uniformly resized to 227 x 227 pixels for consistency. Which is ideal for use in CNNs (Convolutional Neural Networks) and other image-based machine learning models.

Download Dataset

Categories Covered:

Bacterial Spot: This category includes images of tomato leaves infected by the bacterium Xanthomonas campestris, which causes small, water-soaked lesions that can expand and result in tissue necrosis.

Black Leaf Mold: Featuring images of leaves affected by Pseudocercospora fuligena, a fungal disease that produces black spots and mold growth on the underside of leaves.

Gray Leaf Spot: This category captures symptoms of Stemphylium solani infection, characterized by grayish or brownish spots that can lead to leaf desiccation.

Healthy: This class contains images of undiseased tomato leaves, serving as the baseline for comparison against the diseased categories.

Late Blight: A fungal disease caused by Phytophthora infestans, late blight manifests as irregularly shaped lesions with water-soaked margins, often destroying the entire leaf.

Powdery Mildew: Powdery mildew, caused by Oidium neolycopersici, appears as white, powdery patches on the leaves, which can eventually result in chlorosis and leaf drop.

Dataset Features:

Image Diversity: The dataset is rich in visual variation, encompassing images of both singular and multiple leaves, as well as leaves presented against various backgrounds. This diversity helps to mimic real-world conditions where the appearance of leaves can be affected by environmental factors.

Standardized Image Size: To enhance usability in machine learning applications, all images have been resized to a uniform dimension of 227 x 227 pixels, ensuring compatibility with standard deep learning architectures.

Practical Use Cases: This dataset is highly suitable for training and evaluating models in the domains of plant disease classification, agricultural disease prediction, and automated plant health monitoring systems.

Potential Applications:

Agriculture: Supporting Al models that can identify and predict tomato plant diseases early, improving crop yield and reducing the need for manual inspection.

Education: Ideal for use in academic research and projects focused on machine learning, plant pathology, and Al-driven agricultural solutions.

Healthcare for Plants: Assisting farmers and agricultural experts in deploying automated disease detection tools to optimize plant health management.

This dataset is sourced from Kaggle.