DS2Net adds simultaneous detail and semantic deep supervision plus an uncertainty-based adaptive loss and reports consistent but modest gains over state-of-the-art segmentation models on six medical benchmarks.
Nucleus segmentation across imaging experiments: the 2018 data science bowl,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
DS$^2$Net: Detail-Semantic Deep Supervision Network for Medical Image Segmentation
DS2Net adds simultaneous detail and semantic deep supervision plus an uncertainty-based adaptive loss and reports consistent but modest gains over state-of-the-art segmentation models on six medical benchmarks.