CDD: Cucumber Disease Detection Dataset

The original dataset contains a total of 1,280 images, which are divided into eight classes: Anthracnose, Bacterial Wilt, Belly Rot, Downy Mildew, Pythium Fruit Rot, Gummy Stem Blight, Fresh Leaves, and Fresh Cucumber. We divided the dataset into standard proportions for training (70%), validation (20%), and testing (10%). To increase dataset diversity and improve the model’s generalization ability, we used various augmentation techniques, such as horizontal and vertical flipping, rotation from -5° to +5°, horizontal shearing of ±5°, vertical shearing of ±10°, brightness variation from -2% to +2%, and saturation variation from -15% to +15%. We then tripled (3x) the number of training images. After that, the dataset has been expanded to 2839 images.