Abstract

Assessment of fetal head descent is central to intrapartum labour management, but the clinical standard of digital vaginal examination is subjective and examiner-dependent. Transperineal ultrasound enables objective measurement of the angle of progression (AoP), yet manual measurement and automated estimation remain sensitive to image quality, anatomical visibility, and annotation consistency. This study investigates whether dataset curation improves automated AoP estimation. Segmented frames in an intrapartum ultrasound video dataset were quality-classified and their pubic symphysis and fetal head masks manually refined. Segmentation models trained on curated and original masks were compared via downstream AoP, calculated geometrically and referenced against consensus landmark annotations from two experts. Training on a small curated dataset of 356 frames led to a median AoP error of , slightly outperforming models trained on an order of magnitude more non-curated frames, and almost matching the inter-observer variability of . These findings suggest that annotation quality, anatomical consistency, and frame-quality assessment are important for robust automated AoP estimation. We release the curated subset, multi-expert AoP reference, and code as a reproducible benchmark.

Links to Paper and Supplementary Materials

Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/PIPPI_004.pdf

SharedIt Link: Not yet available

SpringerLink (DOI): Not yet available

Supplementary Material: Not Submitted

Link to Open Review

Open Review Page: https://openreview.net/forum?id=jE81y3I36Y&referrer=%5BProgram%20Chair%20Console%5D(%2Fgroup%3Fid%3DMICCAI.org%2F2026%2FWorkshop%2FPIPPI%2FProgram_Chairs%23submission-status)

BibTex

@InProceedings{KalMar_Quality_MICCAISAT2026,
        author = { Kalis, Mareen AND Andelfinger, Nora AND Baumgartner, Christian AND Koch, Lisa M.},
        title = { { Quality over quantity: data curation for automated angle of progression estimation in intrapartum ultrasound } },
        booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
        year = {2026},
        publisher = {Springer Nature Switzerland},
        volume = {LNCS 17257},
        month = {pending},
        page = {pending}
}


back to top