HER2match dataset

This dataset contains 21,172 H&E/HER2 corresponding tile pairs, with 1024x1024 px. The tiles were extracted from 17 pairs of WSI (also included) from breast samples stained with H&E and HER2 (same tissue sample), digitised with a 3DHistech P1000 scanner, at 40X magnification, with a resolution of 0.25 µm/pixel. The tiles dataset is divided into training (11,610 pairs), validation (3,582 pairs), and test (5,980 pairs) sets. More details on acquisition and curation can be found in the corresponding paper. This dataset is part of the project “stAIns: AI-based virtual immuno staining from H&E slides” with file number NGF.1609.241.018 of the research programme NGF AiNED XS Europa 2024, which is (partly) financed by the Dutch Research Council (NWO). If you make use of this dataset, please cite both the dataset itself and the corresponding paper: arXiv:2506.18484