Calibration-Aware Semi-Supervised Fetal Head Segmentation with Boundary-Positive Contrast

Ufaq Khan, Umair Nawaz, Tajamul Ashraf, Tausifa Jan Saleem, Massimo Caputo, Srinivas Ananth Narayan, Muhammad Bilal, Junaid Qadir, Muhammad Haris
Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, PMLR 315:2738-2756, 2026.

Abstract

Accurate fetal head segmentation in ultrasound is hard to scale as labels are scarce and most errors occur at the head–background interface under speckle, shadowing, and low contrast. We present UltraSemiNet, a teacher–student framework that makes cross–pseudo supervision (CPS) selective via temperature calibration and a dual gate requiring high confidence and test-time augmentation (TTA) stability. We also introduce two boundary-focused modules that complement CPS: SAT, a boundary-positive spatial contrast that learns through ambiguous edges using an entropy belt and a soft-IoU agreement test; and PCM, a prototype-guided curriculum that maintains uncertainty-weighted head/background prototypes and targets feature–prototype discrepancies. Across two datasets (FBUI and HC18), UltraSemiNet improves overlap and boundary metrics over a calibrated CPS baseline (e.g., Dice $0.927{\rightarrow}0.971$; HD95 $7.9{\rightarrow}6.8$\,px), with similar cross-dataset trends. Crucially, the calibrated gate reduces miscalibration of the accepted pseudo-labels: both expected calibration error (ECE) and Brier score decrease overall, with the largest gains within the 0–2\,px boundary band, alongside improvements in pseudo-label accuracy. Ablations show CPS calibration, SAT, and PCM are complementary and concentrate improvements on boundary-sensitive metrics. In a blinded study, UltraSemiNet achieved better segmentation performance than two senior fetal medicine experts when evaluated against the dataset reference masks, indicating the potential to reduce manual refinements.

Cite this Paper

BibTeX
@InProceedings{pmlr-v315-khan26a, title = {Calibration-Aware Semi-Supervised Fetal Head Segmentation with Boundary-Positive Contrast}, author = {Khan, Ufaq and Nawaz, Umair and Ashraf, Tajamul and Saleem, Tausifa Jan and Caputo, Massimo and Narayan, Srinivas Ananth and Bilal, Muhammad and Qadir, Junaid and Haris, Muhammad}, booktitle = {Proceedings of The 9th International Conference on Medical Imaging with Deep Learning}, pages = {2738--2756}, year = {2026}, editor = {Huo, Yuankai and Gao, Mingchen and Kuo, Chang-Fu and Jin, Yueming and Deng, Ruining}, volume = {315}, series = {Proceedings of Machine Learning Research}, month = {08--10 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v315/main/assets/khan26a/khan26a.pdf}, url = {https://proceedings.mlr.press/v315/khan26a.html}, abstract = {Accurate fetal head segmentation in ultrasound is hard to scale as labels are scarce and most errors occur at the head–background interface under speckle, shadowing, and low contrast. We present UltraSemiNet, a teacher–student framework that makes cross–pseudo supervision (CPS) selective via temperature calibration and a dual gate requiring high confidence and test-time augmentation (TTA) stability. We also introduce two boundary-focused modules that complement CPS: SAT, a boundary-positive spatial contrast that learns through ambiguous edges using an entropy belt and a soft-IoU agreement test; and PCM, a prototype-guided curriculum that maintains uncertainty-weighted head/background prototypes and targets feature–prototype discrepancies. Across two datasets (FBUI and HC18), UltraSemiNet improves overlap and boundary metrics over a calibrated CPS baseline (e.g., Dice $0.927{\rightarrow}0.971$; HD95 $7.9{\rightarrow}6.8$\,px), with similar cross-dataset trends. Crucially, the calibrated gate reduces miscalibration of the accepted pseudo-labels: both expected calibration error (ECE) and Brier score decrease overall, with the largest gains within the 0–2\,px boundary band, alongside improvements in pseudo-label accuracy. Ablations show CPS calibration, SAT, and PCM are complementary and concentrate improvements on boundary-sensitive metrics. In a blinded study, UltraSemiNet achieved better segmentation performance than two senior fetal medicine experts when evaluated against the dataset reference masks, indicating the potential to reduce manual refinements.} }
Endnote
%0 Conference Paper %T Calibration-Aware Semi-Supervised Fetal Head Segmentation with Boundary-Positive Contrast %A Ufaq Khan %A Umair Nawaz %A Tajamul Ashraf %A Tausifa Jan Saleem %A Massimo Caputo %A Srinivas Ananth Narayan %A Muhammad Bilal %A Junaid Qadir %A Muhammad Haris %B Proceedings of The 9th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2026 %E Yuankai Huo %E Mingchen Gao %E Chang-Fu Kuo %E Yueming Jin %E Ruining Deng %F pmlr-v315-khan26a %I PMLR %P 2738--2756 %U https://proceedings.mlr.press/v315/khan26a.html %V 315 %X Accurate fetal head segmentation in ultrasound is hard to scale as labels are scarce and most errors occur at the head–background interface under speckle, shadowing, and low contrast. We present UltraSemiNet, a teacher–student framework that makes cross–pseudo supervision (CPS) selective via temperature calibration and a dual gate requiring high confidence and test-time augmentation (TTA) stability. We also introduce two boundary-focused modules that complement CPS: SAT, a boundary-positive spatial contrast that learns through ambiguous edges using an entropy belt and a soft-IoU agreement test; and PCM, a prototype-guided curriculum that maintains uncertainty-weighted head/background prototypes and targets feature–prototype discrepancies. Across two datasets (FBUI and HC18), UltraSemiNet improves overlap and boundary metrics over a calibrated CPS baseline (e.g., Dice $0.927{\rightarrow}0.971$; HD95 $7.9{\rightarrow}6.8$\,px), with similar cross-dataset trends. Crucially, the calibrated gate reduces miscalibration of the accepted pseudo-labels: both expected calibration error (ECE) and Brier score decrease overall, with the largest gains within the 0–2\,px boundary band, alongside improvements in pseudo-label accuracy. Ablations show CPS calibration, SAT, and PCM are complementary and concentrate improvements on boundary-sensitive metrics. In a blinded study, UltraSemiNet achieved better segmentation performance than two senior fetal medicine experts when evaluated against the dataset reference masks, indicating the potential to reduce manual refinements.
APA
Khan, U., Nawaz, U., Ashraf, T., Saleem, T.J., Caputo, M., Narayan, S.A., Bilal, M., Qadir, J. & Haris, M.. (2026). Calibration-Aware Semi-Supervised Fetal Head Segmentation with Boundary-Positive Contrast. Proceedings of The 9th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 315:2738-2756 Available from https://proceedings.mlr.press/v315/khan26a.html.

Related Material