Synthetic unfixed diplococci generated microscopic images annotated dataset

This dataset contains 100,000 microscopy images of unfixed diplococci with corresponding binary segmentation masks. The dataset is designed for training, validation, and benchmarking of computer vision models for microscopic image analysis. Each microscopy image is paired with a binary mask highlighting the diplococcal foreground regions. The dataset can be applied to segmentation, image classification, and object detection tasks involving diplococci and other coccoid bacterial structures. Dataset structure The dataset is organized into two folders: images/ — microscopy images in PNG format masks/ — corresponding binary segmentation masks in PNG format Each image and its mask share the same filename. In the masks, white pixels with a value of 255 indicate diplococcal foreground regions, while black pixels with a value of 0 represent the background. Dataset statistics Number of image/mask pairs100,000 File formatPNG