Cervical cell images
This dataset is for the paper titled: Segmentation of Cervical Cell Images based on Generative Adversarial Networks. The dataset is used to train and test the Cell-GAN, a generative adversarial network. After training, the Cell-GAN is able to generate a complete single-cell image which has the similar contour to the cell to be segmented.
The dataset contains two files: a training set, and a test set. For the training set and the test set, each of them is consists of small cell images, annotated cell images, and guide factors. Small cell images are the cropped cell images from the original image set. Guide factors are the annotated nuclei used to locate cell to be segmented. Annotated cell images are the annotated complete single-cell images.
The training set consists of 1571 cell images which includes three types of cells: the upper squamous cells, the mid-level squamous cells, and the basal cells. The test set consists of 309 cell images, including 100 single-cell images, 170 overlapping cell images, and the rest are atypical cell images and test images for ideal cropping. It should be noted that the dataset is used to train and test non-optimized Cell-GAN. So there are same small cell images in each of the two sets, but they correspond to different guide factors.