This dataset is named MSBin which stands for MultiSpectral Document Binarization. The dataset is dedicated to the (document image) binarization of multispectral images. A ReadMe is contained within the dataset and also available here: ReadMe
The dataset is introduced in:
Hollaus, S. Brenner and R. Sablatnig: “CNN based Binarization of MultiSpectral Document Images”. To appear in: International Conference on Document Analysis and Recognition (ICDAR), 2019.