Eggplant_Leaf_Disease_Dataset
Description
Description This dataset, titled Eggplant_Leaf_Disease_Dataset, consists of 3,300 high-resolution images of eggplant (brinjal) leaves classified into six categories. It was developed as part of an academic project by undergraduate students from the Department of Computer Science and Engineering, Netrokona University, Bangladesh. The images are categorized into the following six classes: - Eggplant Healthy Leaf - Eggplant Insect Pest Disease - Eggplant Leaf Spot Disease - Eggplant Mosaic Virus Disease - Eggplant Small Leaf Disease - Eggplant Wilt Disease Each category includes both original images and augmented images. All images are in JPG format and resized to 512x512 pixels using high-quality LANCZOS interpolation to preserve clarity and detail. Data Augmentation To enhance model performance and robustness, the following augmentation techniques were applied using PyTorch’s torchvision library: - Random Horizontal Flip - Rotation (±20 degrees) - Color Jitter (brightness, contrast, saturation, hue) - Random Resized Crop to 512x512 Use Cases This dataset is ideal for: - Image classification tasks -Transfer learning model fine-tuning - Data augmentation research - Academic or student machine learning projects License This dataset is published under the Creative Commons Attribution 4.0 International (CC BY 4.0) license. You can use, share, and adapt the dataset with proper attribution.
Files
Steps to reproduce
Download the Dataset Unzip the dataset after downloading. You will find six folders, each representing a disease class (including healthy leaves). Each folder contains original and augmented images in JPG format. Dataset Structure The folders are organized by class: 1. Eggplant Healthy Leaf 2. Eggplant Insect Pest Disease 3. Eggplant Leaf Spot Disease 4. Eggplant Mosaic Virus Disease 5. Eggplant Small Leaf Disease 6. Eggplant Wilt Disease Image Format - Format: JPG - Size: 512x512 pixels - Color Mode: RGB Augmentation Techniques The dataset includes original and augmented images. The following augmentations were applied using PyTorch’s torchvision.transforms: - RandomHorizontalFlip() - RandomRotation(degrees=20) - ColorJitter(brightness, contrast, saturation, hue) - RandomResizedCrop(size=512) Usage You can use this dataset directly with image classification models (e.g., CNNs) using frameworks like PyTorch or TensorFlow. Simply load the images using standard image loading libraries and split them into training, validation, and test sets as needed. Labeling Labels can be inferred from the folder names for supervised learning.