Segmentation-assisted fetal head descent measurement in intrapartum transperineal ultrasound

Introduction We evaluated whether segmentation of the pubic symphysis and fetal head supports reproducible Angle of Progression (AoP) and Head-Symphysis Distance (HSD) measurement in intrapartum transperineal ultrasound. Methods The PSFHS dataset was partitioned into 951 training, 136 validation, and 271 test images, with exact-image hashes grouped within splits. Nine segmentation architectures were compared during development, and the selected UNet++ workflow was finalized before test evaluation. For each image, AoP and HSD values from three independently trained models were averaged and compared with the arithmetic mean of three senior physicians' measurements. Results AoP mean absolute error was 5.4293 degrees (95% CI, 4.8579-6.0505), and HSD mean absolute error was 2.9904 pixels (95% CI, 2.5915-3.4285). Individual-model MAEs ranged from 5.7808 to 5.9612 degrees for AoP and from 3.1931 to 3.3237 pixels for HSD. Foreground Dice was 0.9449 ± 0.0011 across training runs, with no geometric measurement failures on the test images. A validation ablation showed that a small gain in segmentation overlap could accompany larger measurement errors. Discussion These results provide internal technical validation of the segmentation-to-measurement pathway against a physician mean reference. HSD remains in pixels, and patient-level independence could not be verified. External validation, reliable physical calibration, and prospective evaluation are needed to establish the workflow's usefulness in clinician-led intrapartum assessment.

Authors

Institutions

Publication Details

Journal
Frontiers in Medicine
Published
2026-11-18
DOI
https://doi.org/10.3389/fmed.2026.1921165
Primary Topic
Medical Image Segmentation Techniques
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Segmentation-assisted fetal head descent measurement in intrapartum transperineal ultrasound

Jing Qiu, 杨之源, Linlin Song, Dongren Liu et al.
3 links
Frontiers in Medicine
Medical Image Segmentation Techniques
article

Segmentation-assisted fetal head descent measurement in intrapartum transperineal ultrasound

Jing Qiu, 杨之源, Linlin Song, Dongren Liu, Daozhi Qian, Bentao Jiang, Shuqing Li
article en

Abstract

Introduction We evaluated whether segmentation of the pubic symphysis and fetal head supports reproducible Angle of Progression (AoP) and Head-Symphysis Distance (HSD) measurement in intrapartum transperineal ultrasound. Methods The PSFHS dataset was partitioned into 951 training, 136 validation, and 271 test images, with exact-image hashes grouped within splits. Nine segmentation architectures were compared during development, and the selected UNet++ workflow was finalized before test evaluation. For each image, AoP and HSD values from three independently trained models were averaged and compared with the arithmetic mean of three senior physicians' measurements. Results AoP mean absolute error was 5.4293 degrees (95% CI, 4.8579-6.0505), and HSD mean absolute error was 2.9904 pixels (95% CI, 2.5915-3.4285). Individual-model MAEs ranged from 5.7808 to 5.9612 degrees for AoP and from 3.1931 to 3.3237 pixels for HSD. Foreground Dice was 0.9449 ± 0.0011 across training runs, with no geometric measurement failures on the test images. A validation ablation showed that a small gain in segmentation overlap could accompany larger measurement errors. Discussion These results provide internal technical validation of the segmentation-to-measurement pathway against a physician mean reference. HSD remains in pixels, and patient-level independence could not be verified. External validation, reliable physical calibration, and prospective evaluation are needed to establish the workflow's usefulness in clinician-led intrapartum assessment.

Frontiers in MedicineVol. 13
Qingdao University (CN), Tianjin University of Sport (CN), Shandong Maternal and Child Health Hospital (CN), Jinan Third People's Hospital (CN)
Good health and well-being
Medical Image Segmentation Techniques
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon