what is the difference between the two training code(certainty/certainty_unlabel)

yulequan / UA-MT

code for MICCAI 2019 paper 'Uncertainty-aware Self-ensembling Model for Semi-supervised 3D Left Atrium Segmentation'.

https://arxiv.org/abs/1907.07034

477 stars 97 forks source link

Closed yykhuster closed 3 years ago

yulequan commented 3 years ago

The difference is whether we add the consistency loss on labeled data.