Using the network's previous epoch as a teacher, with patch-level KL divergence and pixel-wise soft labels, improves coronary artery segmentation Dice by 2 to 4 points on XCAD and DCA1.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
eess.IV 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Deep Self-knowledge Distillation: A hierarchical supervised learning for coronary artery segmentation
Using the network's previous epoch as a teacher, with patch-level KL divergence and pixel-wise soft labels, improves coronary artery segmentation Dice by 2 to 4 points on XCAD and DCA1.