Training data resolution

Dear organizers, the data is described as TIF files at 0.5 um/px. However I am confused by the submission guids at the "submission requirements" page. It says that "To convert the slide-level (pixel) coordinates to mm: coordspacing[um]/1000." and I believe this conversion is for resolutions of 1.0um/px. Am I understanding wrong or is it because the test data are images scanned at a different magnification. Thank you.

Sincerely, krlee

Dear Kyung Ryun Lee,

The division by 1000 is required because the grand-challenge platform expects to coordinate to be in mm instead of um. Please also note that the grand challenge expects the coordinates to be in 'world coordinates.' Therefore you will also have to multiply them with the spacing value. Please see an example here: https://github.com/DIAGNijmegen/pathology-tiger-algorithm-example/blob/b24bb2e075fe60964278515aa8e125a6162037bf/tigeralgorithmexample/rw.py#L119

Best wishes, Mart