Lemon Leaf Diseases Dataset
Description
This dataset comprises 3,100 images, evenly categorized into 4 distinct classes. It was developed for a computer vision project undertaken by students in the Department of Computer Science and Engineering. Netrokona University, Netrokona, Bangladesh. The collection features Original Images, all uniformly resized to 512x512 pixels using high-quality LANCZOS interpolation to maintain visual clarity. ## Folder Structure The dataset is organized into the following class folders: -Canker -Black Spot -Healthy -Huanglongbing Original images: -Canker(775) -Black Spot(775) -Healthy(775) -Huanglongbing(775)
Files
Steps to reproduce
## Data Augmentation Techniques To increase variability and robustness for machine learning models, the following augmentations were applied: - Random Horizontal Flip - Rotation up to ±20 degrees - Color Jitter (brightness, contrast, saturation, hue) - Random Resized Crop to 512x512 These transformations were applied using PyTorch's torchvision library. ## Image Format - Format: JPG - Size: 512x512 pixels - Color Mode: RGB ## Use Case This dataset is suitable for: - Image classification tasks - Transfer learning model fine-tuning - Federated Learning - Dataset augmentation research - Student or academic thesis
Institutions
- Netrokona UniversityMymensingh Division, Netrakona