Training & Testing Data

From the stage ONE, 150 samples with reference standards are released as training data. The remaining 50 samples are test data. Participants are allowed to use their own training data as well. All data will be uploaded to Microsoft OneDrive and Baidu Pan for participants from different regions. Whole-Slide images are in TIFF format, manual annotations are in XML format.

In the clinical practice, more than one sample from the same biopsy will be scanned. If samples have a similar shape, the pathologist only annotated one sample for the WSI (see the figure below, Patient NO.60). Participants can ignore other samples during the training stage.

We suggest participants use ASAP to make a bounding box themselves to exclude the unused samples. Please see the RGB arrows in the figure. RGB corresponds to rectangle annotation button that makes bounding box, rectangle annotation name (you can change "Annotation 7" into "boundingbox" or other name you prefer to) and the bounding box in the image.