Search

This the part 1 of our deidentified dataset that used in our article: BCR-Net: A deep learning framework to predict breast cancer recurrence from histopathology images. Our complete dataset contains the whole slide images (WSIs) of 151 H&E breast cancer resection tissues. For each WSI, we extracted the foreground ROIs and cropped it into 224x224 patches under 40x magnification. Then, we save the cropped images into H5 files. In each H5 file, we saved the extracted patches from the WSI in the "bag" subset, and the patches' corresonding coordinates on the original WSI in the "coords" subset. We saved the ODX scores of the WSIs in the Labels.xlsx. All the data are deidentified. &nbsp

This the part 2 of our deidentified dataset that used in our article: BCR-Net: A deep learning framework to predict breast cancer recurrence from histopathology images. Our complete dataset contains the whole slide images (WSIs) of 151 H&E breast cancer resection tissues. For each WSI, we extracted the foreground ROIs and cropped it into 224x224 patches under 40x magnification. Then, we save the cropped images into H5 files. In each H5 file, we saved the extracted patches from the WSI in the "bag" subset, and the patches' corresonding coordinates on the original WSI in the "coords" subset. We saved the ODX scores of the WSIs in the Labels.xlsx. All the data are deidentified.