You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Thanks for the question. The HD95 value would be nan if there are no foregrounds in either the prediction or label. It was set to be ~373 following the BraTS challenge (math.sqrt(depth2 + width2 + height**2)), which is not exact but only a penalization item. Another alternative solution is ignoring the cases with nan HD95, which may lead to lower HD.
I noticed that when the number of nonzero voxels is zero, the hd95 is artificially set to 373.12866. Why choose it not other?
The text was updated successfully, but these errors were encountered: