Cell Nuclei Classification Dataset

Dataset description Cell Nuclei Classification Dataset is comprised of 63,228 images crops, captured in a Sartorius IncuCyte Incubator/Microscope, with 20x magnification. This dataset was generated in order to train a deep learning model for nuclei classification in fluorescence/brightfield images (unpublished) at LabSinal laboratory (https://www.ufrgs.br/labsinal/), and contains annotations on two phenotypes: autophagy and cell cycle. Annotations are described in the "dataset_df.csv" in each respective subfolder, and are essentially described by columns: crop_name: represents the file name stored in "train", "val" or "test" subfolders. treatment: stands for the treatment group of the original image (CTR: control, TMZ: temozolomide) class: represents the label/phenotype of the given nucleus (which was obtained using the information from the fluorescence channel of the respective nucleus) split: the split group (dataset is balanced, and splits are stratified according to the class and treatment, to avoid learning bias)