A contrastive self-supervised method with frequency and patch augmentations and a learned relation loss improves breast ultrasound segmentation and cross-dataset transfer, especially with little labelled data.
Cross-Image Dependency Modeling for Breast Ultrasound Segmentation,
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
-
A Self-Supervised Framework for Improved Generalisability in Ultrasound B-mode Image Segmentation
A contrastive self-supervised method with frequency and patch augmentations and a learned relation loss improves breast ultrasound segmentation and cross-dataset transfer, especially with little labelled data.