gzuidhof/luna16

What is the input of "false positive reduction" network?

Bonsen opened this issue · 1 comments

What is the input of "false positive reduction" network?

After the prediction of unet, u merge the candidates and then extract patches according to the candidate x,y,z with label 1/0? Or u just extract patches from original images according the x,y,z label for training and then predict the candidate?