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

Khan, Ufaq and Nawaz, Umair and Ashraf, Tajamul and Saleem, Tausifa and Caputo, Massimo and Narayan, Srinivas Anath and Bilal, Muhammad and Qadir, Junaid and Haris, Muhammad (2026) Calibration-Aware Semi-Supervised Fetal Head Segmentation with Boundary-Positive Contrast. In: Medical Imaging with Deep Learning, 8th-10th July 2026, Chientan, Taipei, Taiwan.

Preview

Text khan26a.pdf - Published Version Available under License Creative Commons Attribution. Download (3MB)

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→0.971; HD95 7.9→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.

Item Type:	Conference or Workshop Item (Paper)
Dates:	Date Event 1 July 2026 Accepted 10 July 2026 Published Online
Uncontrolled Keywords:	Fetal Segmentation, Semi-Supervised Learning, Boundary-Aware Learning
Subjects:	CAH17 - business and management > CAH17-01 - business and management > CAH17-01-01 - business and management (non-specific)
Divisions:	Business School > Management, Business and Marketing
Depositing User:	Gemma Tonks
Date Deposited:	04 Sep 2026 14:19
Last Modified:	04 Sep 2026 14:19
URI:	https://www.open-access.bcu.ac.uk/id/eprint/17214

Actions (login required)

View Item