GrEp_Dataset

################################################################################## #------------------------------------ Overview ----------------------------------------------------------------------- ################################################################################## This dataset was first used and released together with the paper: GrEp: Graph-based epithelial cell classification refinement in histopathology H&E images If you use the dataset or the corresponding code, please cite the paper accordingly. @article{frei2025, title={GrEp: Graph-based epithelial cell classification refinement in histopathology H&E images}, author={Frei, Ana Leni and Garcia-Baroja, Javier and Rau, Tilman and Neppl, Christina and Lugli, Alessandro and Solass, Wiebke and Wartenberg, Martin and Fischer, Andreas and Zlobec, Inti}, journal={Journal of Pattern Recognition}, year={2025}, publisher={Elsevier}} ################################################################################## #------------------------------- Dataset Description ----------------------------------------------------------------- ################################################################################## The dataset is composed of 3 sub-datasets: 1) Colorectal epithelium 2) Endometrium epithelium 3) Pancreatic epithelium All tiles are provided at 20X magnification (0.5micrometer/pixel). The annotations corresponding to each dataset are organized following the same logic. To each tile corresponds a .mat file, where: - 'inst_centroid: Nx2 array, (x,y) coordinates of the nuclei, where N corresponds to the number of cells in the tile - 'inst_type': Nx1 array, nuclei label with: 0 = normal epithelium 1 = malignant epithelium 2 = other