Jiabang/Single-cell_Microscopy_Images_for_Cancer_Classification

This datatset is a mirror to the dataset "Single-cell Microscopy Images for Cancer Classification" in Mendeley Data, we will maintain both to upload new dataset in time and keep these two versions the same.

This is a dataset for single cell fluorescence microscopy images. The dataset includes cancer cells and normal cells. The images contain three channels: red (actin filament), green (vimentin filament), and blue (nucleus). We have 591 normal cells and 542 cancer cells. More information about how to culture the cells, stain the cells, take the photos, and cell segmentation can be found in our paper "A new explainable AI pipeline for cancer cell segmentation and classification for microscopy images". Please cite this paper if you find our dataset is useful.

This is a well-labelled and continuously updated dataset, we plan to upload more cell images, not only just for cancer and normal cells, but also upload some cells which are in the stages between normal and cancer, which will provide other two classes. Besides, we will also upload cell images contain another channel: microtube.

Citation: The paper has been accepted by BMC Bioinformatics, please cite: @article{chen2026new, title={A new explainable AI pipeline for cancer cell segmentation and classification for microscopy images}, author={Chen, Jiabang and Kezi{'c}, Antonija and Alwadiya, Ameer and Gad, Annica KB and Panoutsos, George}, journal={BMC Bioinformatics}, year={2026}, publisher={Springer} } DOI: https://doi.org/10.1186/s12859-026-06680-0

Contact: Jiabang Chen: jchen144@sheffield.ac.uk