ET-Net adds an edge guidance module that extracts edge-attention maps in early layers and fuses them via weighted aggregation in the decoder, improving segmentation accuracy on optic disc/cup, vessel, and lung tasks over prior methods.
IEEE TMI (2004)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2019 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
ET-Net: A Generic Edge-aTtention Guidance Network for Medical Image Segmentation
ET-Net adds an edge guidance module that extracts edge-attention maps in early layers and fuses them via weighted aggregation in the decoder, improving segmentation accuracy on optic disc/cup, vessel, and lung tasks over prior methods.