Histological images for tumor detection in gastrointestinal cancer

This is a set of 11977 image patches of hematoxylin & eosin stained histological samples of human colorectal cancer. It is a subset of the data set "100,000 histological images of human colorectal cancer and healthy tissue" which is accessible at http://dx.doi.org/10.5281/zenodo.1214456. Further information on the samples is available there.

This set contains three classes:

ADIMUC - adipose tissue and mucus, i.e. loose non-tumor tissue

STRMUS - stroma and muscle, i.e. dense non-tumor tissue

TUMSTU - colorectal cancer epithelial tissue and stomach cancer epithelial tissue, i.e. tumor tissue

All images are 512x512 px at 0.5 µm/px

We are using this data set to train a deep neural network to detect tumor cells in histological whole slide images of colorectal and stomach cancer.

If you use these images in your research, please consider citing our previous [1] and upcoming publication.

For information on ethics board approval, see [1].

-------------

[1] http://dx.doi.org/10.1371/journal.pmed.1002730