R2AoP: Reliable and Robust Angle of Progression Estimation from Intrapartum Ultrasound

Beining Wu; Changmiao Wang; Chunbo Jiang; Feiwei Qin; Mingxuan Liu; Qiyuan Tian; Xiaotian Hu; Yifei Chen; Yijin Li; Yuanhan Wang

arxiv: 2605.21099 · v1 · pith:4ZUKBCCKnew · submitted 2026-05-20 · 💻 cs.CV

R2AoP: Reliable and Robust Angle of Progression Estimation from Intrapartum Ultrasound

Yuanhan Wang , Yifei Chen , Beining Wu , Mingxuan Liu , Xiaotian Hu , Chunbo Jiang , Yijin Li , Changmiao Wang

show 2 more authors

Feiwei Qin Qiyuan Tian

This is my paper

Pith reviewed 2026-05-21 05:09 UTC · model grok-4.3

classification 💻 cs.CV

keywords Angle of ProgressionIntrapartum UltrasoundSegmentationTest-time AdaptationFetal HeadPubic SymphysisMedical Image AnalysisLabor Progression

0 comments

The pith

R2AoP combines a three-branch segmentation backbone with confidence-weighted contour fitting and test-time adaptation to produce more stable Angle of Progression measurements from intrapartum ultrasound.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops R2AoP to address the sensitivity of Angle of Progression estimates to ultrasound noise and boundary ambiguities that arise during labor assessment. It employs a specialized backbone to better outline the pubic symphysis and fetal head, then applies confidence scores to down-weight uncertain contour segments when computing the angle. A lightweight adaptation step further stabilizes results when image acquisition varies across centers. A sympathetic reader would care because more reproducible measurements could support consistent clinical decisions about labor progress without requiring fresh annotations for each new scanner or protocol.

Core claim

R2AoP integrates structurally informed segmentation and confidence-guided geometric modeling to achieve stable and reproducible AoP measurements. A three-branch local-structure-enhanced backbone improves the delineation of the pubic symphysis and fetal head, while confidence-weighted contour fitting explicitly suppresses the influence of unreliable boundary points in AoP computation. A lightweight geometry-reliable test-time adaptation strategy enables stable inference under heterogeneous acquisition conditions without target annotations, yielding consistent reductions in AoP error and boundary metrics on multi-center benchmarks.

What carries the argument

Confidence-weighted contour fitting, which down-weights low-confidence boundary points when deriving the angle between the pubic symphysis and fetal head contours.

If this is right

Clinicians could obtain more reproducible labor-progress assessments across different operators and devices.
Fewer manual boundary corrections would be needed in routine intrapartum ultrasound workflows.
Research studies using AoP as an outcome measure would face lower measurement variability.
The same weighting and adaptation components could be reused for other geometric fetal measurements in ultrasound.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

Similar confidence-guided fitting could reduce error amplification in other noisy geometric tasks such as vessel-angle measurements in echocardiography.
If the adaptation generalizes, the method might serve as a template for domain-shift handling in additional obstetric imaging applications without new labeled data.
Clinical outcome studies could test whether the observed metric improvements correlate with changes in delivery decisions or cesarean rates.

Load-bearing premise

That confidence scores from the segmentation model will correctly identify unreliable boundary points and that the test-time adaptation will not create new instabilities when applied to acquisition conditions absent from the training data.

What would settle it

Performance measurements on a fresh multi-center ultrasound dataset acquired with scanners or patient demographics outside the original training distribution, checking whether AoP error reductions and boundary metric improvements hold or reverse.

Figures

Figures reproduced from arXiv: 2605.21099 by Beining Wu, Changmiao Wang, Chunbo Jiang, Feiwei Qin, Mingxuan Liu, Qiyuan Tian, Xiaotian Hu, Yifei Chen, Yijin Li, Yuanhan Wang.

**Figure 1.** Figure 1: Illustration of the source-to-target pipeline of the proposed method. Sourcedomain training is followed by target-domain inference with geometry-reliable feedback for AoP computation. AoP is determined by the pubic symphysis axis and the tangent to the fetal head contour [5], automated computation requires robust PS/FH localization and accurate segmentation [21]. However, as a cascaded geometric measure,… view at source ↗

**Figure 2.** Figure 2: The overall framework of the proposed method R2AoP. [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

**Figure 3.** Figure 3: Qualitative comparison of PS/FH segmentation and AoP estimation across methods. The top row shows segmentation results, and the bottom row shows AoP geometry visualization including the PS axis and FH tangent [PITH_FULL_IMAGE:figures/full_fig_p007_3.png] view at source ↗

**Figure 4.** Figure 4: Qualitative ablation results of different components in the proposed method. The left panels show segmentation error maps with TP, FP, and FN, and the right panels show confidence maps with AoP error and boundary metrics. 3.2 Experimental details All experiments are conducted on a single NVIDIA A800 GPU, with the software environment consisting of Ubuntu 20.04.6 and Python 3.9. During source-domain trainin… view at source ↗

read the original abstract

Accurate estimation of the Angle of Progression (AoP) from intrapartum transperineal ultrasound is critical for objective assessment of labor progression, yet remains highly sensitive to imaging noise, boundary ambiguities, and the geometric amplification of local segmentation errors. We propose R2AoP, a reliable and robust AoP estimation framework that integrates structurally informed segmentation and confidence-guided geometric modeling to achieve stable and reproducible measurements. A three-branch local-structure-enhanced backbone improves the delineation of the pubic symphysis (PS) and fetal head (FH), while confidence-weighted contour fitting explicitly suppresses the influence of unreliable boundary points in AoP computation. To further improve performance under heterogeneous acquisition conditions, we introduce a lightweight geometry-reliable test-time adaptation strategy as an auxiliary component, enabling stable inference without target annotations. Extensive evaluations on multi-center benchmarks demonstrate consistent reductions in AoP error and boundary metrics compared with state-of-the-art AoP methods. Our source code is available at https://github.com/baiyou1234/R2AoP.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

R2AoP assembles a three-branch backbone, confidence-weighted fitting, and test-time adaptation for AoP estimation in a sensible way for this clinical task, but the adaptation's claimed robustness needs direct evidence from ablations and shift tests.

read the letter

Hi there, The one or two things to know about this paper are that it introduces R2AoP, which integrates a three-branch backbone designed to capture local structures better for segmenting the pubic symphysis and fetal head in intrapartum ultrasound, uses confidence-weighted contour fitting to minimize the effect of noisy or ambiguous boundary points when calculating the angle of progression, and employs a lightweight test-time adaptation strategy to enforce geometric reliability during inference on new data without annotations. The authors claim consistent improvements in AoP error and boundary metrics across multi-center datasets, and they've shared the source code on GitHub. What the paper does well is focus on a real clinical pain point. AoP estimation is important for assessing labor but sensitive to imaging issues, and the method tries to make it more stable by combining segmentation improvements with geometric modeling. The test-time adaptation as an auxiliary component is a practical touch for handling heterogeneous conditions. Providing the code allows others to verify or extend the work, which is good practice. Where it might be soft is in the validation of the adaptation part. The stress-test concern about whether the geometric prior holds under acquisition shifts not seen in training is worth paying attention to. Without detailed ablations showing performance with and without the adaptation, or tests on deliberately varied conditions, it's hard to be sure the reported reductions aren't partly due to overfitting to the training distribution. The abstract doesn't include quantitative results, so the full paper's results section will be key to assessing if the evidence supports the claims strongly. This paper is for specialists in medical image analysis and obstetrics imaging. A reader working on similar problems in ultrasound or needing ideas for robust geometric estimation from segmentation could get some value from the specific design choices and the benchmark comparisons. I think it deserves a serious referee because it's addressing a targeted but important application with a clear method, even if revisions might be needed for stronger evidence on generalization.

Referee Report

2 major / 2 minor

Summary. The manuscript proposes R2AoP, a framework for reliable Angle of Progression (AoP) estimation from intrapartum transperineal ultrasound. It integrates a three-branch local-structure-enhanced backbone for delineating the pubic symphysis (PS) and fetal head (FH), confidence-weighted contour fitting to suppress unreliable boundary points, and a lightweight geometry-reliable test-time adaptation strategy (without target annotations) to handle heterogeneous acquisition conditions. The central claim is that these components yield stable, reproducible AoP measurements with consistent reductions in AoP error and boundary metrics versus state-of-the-art methods on multi-center benchmarks.

Significance. If the robustness claims hold under the reported conditions, the work could meaningfully improve objective intrapartum assessment by mitigating sensitivity to noise and boundary ambiguity. The annotation-free test-time adaptation and open-source code are practical strengths that support reproducibility and potential clinical translation in medical imaging.

major comments (2)

[§3.3] §3.3 (test-time adaptation): The geometry-reliable adaptation is presented as key to robustness under heterogeneous conditions, yet no ablation isolates its effect on AoP error when acquisition parameters (probe angle, depth, or patient anatomy) are deliberately shifted outside the multi-center training distribution. Without this, the reported reductions cannot be confidently attributed to genuine generalization rather than reinforcement of training-center priors.
[§4.2] §4.2 and Table 3: The multi-center results claim consistent AoP error reductions, but the manuscript supplies neither per-center breakdowns with error bars nor statistical tests for significance of the improvements; this weakens the cross-center reproducibility assertion that underpins the central claim.

minor comments (2)

[Abstract] Abstract: The phrase 'consistent reductions in AoP error and boundary metrics' is stated without numerical magnitudes or the precise metrics (e.g., mean absolute error, Hausdorff distance), reducing immediate clarity for readers.
[§2.1] §2.1: The definition of the confidence-weighted contour fitting could benefit from an explicit equation showing how unreliable points are down-weighted in the AoP angle computation.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive feedback on our manuscript. We address each major comment below and outline the revisions we will make to strengthen the presentation of our results.

read point-by-point responses

Referee: [§3.3] §3.3 (test-time adaptation): The geometry-reliable adaptation is presented as key to robustness under heterogeneous conditions, yet no ablation isolates its effect on AoP error when acquisition parameters (probe angle, depth, or patient anatomy) are deliberately shifted outside the multi-center training distribution. Without this, the reported reductions cannot be confidently attributed to genuine generalization rather than reinforcement of training-center priors.

Authors: We agree that an explicit ablation isolating the test-time adaptation under controlled out-of-distribution shifts would provide stronger evidence for its contribution. Our current multi-center results already reflect performance across varied acquisition settings, but they do not include deliberate parameter perturbations. In the revised manuscript we will add a targeted ablation study that simulates shifts in probe angle and depth outside the training distribution and reports the resulting changes in AoP error. revision: yes
Referee: [§4.2] §4.2 and Table 3: The multi-center results claim consistent AoP error reductions, but the manuscript supplies neither per-center breakdowns with error bars nor statistical tests for significance of the improvements; this weakens the cross-center reproducibility assertion that underpins the central claim.

Authors: We acknowledge that per-center breakdowns and formal statistical tests are necessary to support the reproducibility claims. The current Table 3 aggregates results across centers without these details. In the revised version we will expand Section 4.2 and Table 3 to include per-center mean AoP errors with standard-deviation error bars and will report statistical significance (e.g., paired Wilcoxon signed-rank tests) for the observed improvements over baselines. revision: yes

Circularity Check

0 steps flagged

No derivation chain present; empirical framework evaluated on external benchmarks

full rationale

The manuscript describes an engineering pipeline (three-branch backbone, confidence-weighted contour fitting, lightweight test-time adaptation) whose performance claims rest on multi-center empirical benchmarks rather than any closed-form derivation or first-principles prediction. No equations appear that define a quantity in terms of itself or that rename a fitted parameter as a 'prediction.' Self-citations, if present, are not load-bearing for the central robustness claim, which is supported by external data splits. The derivation chain is therefore self-contained and non-circular.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

Only the abstract is available, so the ledger is necessarily incomplete; no explicit free parameters, axioms, or invented entities are stated beyond the high-level architectural choices.

axioms (1)

domain assumption A three-branch local-structure-enhanced backbone improves delineation of the pubic symphysis and fetal head.
Invoked as the basis for the segmentation stage in the abstract.

pith-pipeline@v0.9.0 · 5742 in / 1286 out tokens · 32089 ms · 2026-05-21T05:09:04.884161+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

24 extracted references · 24 canonical work pages

[1]

Bai, J., Lekadir, K., Ni, D., Slimani, S., Campello, V.M., Ohene-Botwe, B., Lu, Y., Chen, G., Hou, H., Qiu, D., Zhou, Z.: Intrapartum ultrasound grand challenge 2024 (2024)

work page 2024
[2]

Medical Image Analysis99, 103353 (2025)

Bai, J., Zhou, Z., Ou, Z., Koehler, G., Stock, R., Maier-Hein, K., Elbatel, M., Martí, R., Li, X., Qiu, Y., Gou, P., Chen, G., Zhao, L., Zhang, J., Dai, Y., Wang, F., Silvestre, G., Curran, K., Sun, H., Xu, J., Cai, P., Jiang, L., Lan, L., Ni, D., Zhong, M., Chen, G., Campello, V.M., Lu, Y., Lekadir, K.: PSFHS challenge report: Pubic symphysis and fetal h...

work page 2025
[3]

Medical Image Analysis101, 103420 (2025)

Campello, V.M., Lekadir, K., Bai, J., et al.: Automated angle of progression es- timation from intrapartum ultrasound: A multi-center benchmark study. Medical Image Analysis101, 103420 (2025)

work page 2025
[4]

Scientific Data11(1), 436 (2024)

Chen, G., Bai, J., Ou, Z., Lu, Y., Wang, H.: PSFHS: intrapartum ultrasound image dataset for AI-based segmentation of pubic symphysis and fetal head. Scientific Data11(1), 436 (2024)

work page 2024
[5]

In: Proceedings of the 6th ACM International Conference on Multimedia in Asia

Chen, S., Wang, H., Long, S., Bai, J., Jiang, J.: Ultrasound video segmentation of pubic symphysis and fetal head for AoP measurement. In: Proceedings of the 6th ACM International Conference on Multimedia in Asia. MMAsia ’24, Association for Computing Machinery, New York, NY, USA (2024)

work page 2024
[6]

Computers in Biology and Medicine170, 107917 (2024) 10 Y

Chen, Y., Zhang, C., Chen, B., Huang, Y., Sun, Y., Wang, C., Fu, X., Dai, Y., Qin, F., Peng, Y., Gao, Y.: Accurate leukocyte detection based on deformable-DETR and multi-level feature fusion for aiding diagnosis of blood diseases. Computers in Biology and Medicine170, 107917 (2024) 10 Y. Wang et al

work page 2024
[7]

IEEE Journal of Biomedical and Health Infor- matics28(8), 4648–4659 (2024)

Chen, Z., Lu, Y., Long, S., Campello, V.M., Bai, J., Lekadir, K.: Fetal head and pubic symphysis segmentation in intrapartum ultrasound image using a dual-path boundary-guided residual network. IEEE Journal of Biomedical and Health Infor- matics28(8), 4648–4659 (2024)

work page 2024
[8]

Expert Systems with Applications296, 128998 (2026)

Chen, Z., Ou, Z., Lu, Y., Campello, V.M., Bai, J., Lekadir, K.: Uncertainty- fetal head and pubic symphysis segmentation with enhanced multi-scale features and sparse visual graph attention. Expert Systems with Applications296, 128998 (2026)

work page 2026
[9]

In: Bai, J., Lu, Y

Gan, J., Liang, Z., Fan, J., Mcguire, L., Clarke, J., Cai, W.: Accurate fetal head descent assessment during labor using video Swin Transformer and wavelet-based multitask learning for 2024 MICCAI challenge IUGC. In: Bai, J., Lu, Y. (eds.) Intrapartum Ultrasound. pp. 21–31. Cham (2026)

work page 2024
[10]

In: Linguraru, M.G., Dou, Q., Feragen, A., Giannarou, S., Glocker, B., Lekadir, K., Schnabel, J.A

Jiang, J., Wang, H., Bai, J., Long, S., Chen, S., Campello, V.M., Lekadir, K.: In- trapartum Ultrasound Image Segmentation of Pubic Symphysis and Fetal Head Using Dual Student-Teacher Framework with CNN-ViT Collaborative Learning. In: Linguraru, M.G., Dou, Q., Feragen, A., Giannarou, S., Glocker, B., Lekadir, K., Schnabel, J.A. (eds.) Medical Image Comput...

work page 2024
[11]

Medical Image Analysis96, 103202 (2024)

Jiao, J., Zhou, J., Li, X., Xia, M., Huang, Y., Huang, L., Wang, N., Zhang, X., Zhou, S., Wang, Y., Guo, Y.: USFM: A universal ultrasound foundation model generalized to tasks and organs towards label efficient image analysis. Medical Image Analysis96, 103202 (2024)

work page 2024
[12]

In: Bai, J., Huang, Y., Khobo, I., Yaqub, M., Lekadir, K., Ni, D., Li, S

Liu, X., Hu, J., Li, Y., Chen, X., Wang, Y.: Noisy student-based self-training enhances landmark detection in intrapartum ultrasound. In: Bai, J., Huang, Y., Khobo, I., Yaqub, M., Lekadir, K., Ni, D., Li, S. (eds.) Intrapartum Ultrasound. pp. 1–13. Cham (2026)

work page 2026
[13]

Data in Brief41, 107904 (2022)

Lu, Y., Zhou, M., Zhi, D., Zhou, M., Jiang, X., Qiu, R., Ou, Z., Wang, H., Qiu, D., Zhong, M., Lu, X., Chen, G., Bai, J.: The JNU-IFM dataset for segmenting pubic symphysis-fetal head. Data in Brief41, 107904 (2022)

work page 2022
[14]

Medical Image Analysis109, 103902 (2026)

Ma, X., Tao, Y., et al.: Test-time generative augmentation for medical image seg- mentation. Medical Image Analysis109, 103902 (2026)

work page 2026
[15]

Computers in Biology and Medicine175, 108501 (2024)

Ou, Z., Bai, J., Chen, Z., Lu, Y., Wang, H., Long, S., Chen, G.: RTSeg-net: A lightweight network for real-time segmentation of fetal head and pubic symphysis from intrapartum ultrasound images. Computers in Biology and Medicine175, 108501 (2024)

work page 2024
[16]

IEEE Transactions on Medical Imaging44(6), 2553–2567 (2025)

Ravishankar, H., Paluru, N., Sudhakar, P., Yalavarthy, P.K.: Information Geomet- ric approaches for patient-specific test-time adaptation of deep learning models for semantic segmentation. IEEE Transactions on Medical Imaging44(6), 2553–2567 (2025)

work page 2025
[17]

In: Bai, J., Lu, Y

Tan, Y., Wang, S., Shen, W.D., Yu, W.W., Li, Y., Zhang, P., Jiang, W., Li, Y.J.: Classification and segmentation of intrapartum ultrasound images with deep learn- ing models. In: Bai, J., Lu, Y. (eds.) Intrapartum Ultrasound. pp. 11–20. Cham (2026)

work page 2026
[18]

In: Intrapartum Ultrasound: MICCAI 2025 Grand Challenge, IUGC 2025, Held in Conjunction with MICCAI 2025, Daejeon, South Korea, September 23, 2025, Proceedings

Tang, Y., Zhou, Z., Lu, Y., Bai, J., Long, S., Huang, Y., Khobo, I., Zhang, S., Zhou, Z., Guo, L.: Heatmap regression for automated angle of progression mea- surement: The baseline method for the IUGC2025. In: Intrapartum Ultrasound: MICCAI 2025 Grand Challenge, IUGC 2025, Held in Conjunction with MICCAI 2025, Daejeon, South Korea, September 23, 2025, Pro...

work page 2025
[19]

Medical Physics 52(4), 2311–2324 (2025)

Wang, H., Chen, S., Bai, J.: Temporal modeling of fetal head descent for robust angle of progression estimation in intrapartum ultrasound videos. Medical Physics 52(4), 2311–2324 (2025)

work page 2025
[20]

In: Bai, J., Lu, Y

Xia, Z., Li, H., Lan, L.: DSSAU-Net: U-shaped hybrid network for pubic symphysis and fetal head segmentation. In: Bai, J., Lu, Y. (eds.) Intrapartum Ultrasound. pp. 32–45. Cham (2026)

work page 2026
[21]

Zhou, M., Lu, Y., Bai, J.: Direct regression of the angle of progression in intra- partumultrasoundusingdeepneuralnetworks.ComputersinBiologyandMedicine 173, 108312 (2024)

work page 2024
[22]

Expert Systems with Applications263, 125699 (2025)

Zhou, Z., Lu, Y., Bai, J., Campello, V.M., Feng, F., Lekadir, K.: Segment Anything Model for fetal head-pubic symphysis segmentation in intrapartum ultrasound im- age analysis. Expert Systems with Applications263, 125699 (2025)

work page 2025
[23]

Medical Image Analysis112, 104108 (2026)

Zhu, S., Chen, Y., Chen, W., Jiang, S., Zhou, G., Wang, Y., Qin, F., Wang, C., Tian, Q.: No modality left behind: Adapting to missing modalities via knowledge distillation for brain tumor segmentation. Medical Image Analysis112, 104108 (2026)

work page 2026
[24]

In: proceedings of Medical Image ComputingandComputerAssistedIntervention–MICCAI2025.vol.LNCS15967

Zhu, S., Chen, Y., Chen, W., Wang, Y., Liu, C., Jiang, S., Qin, F., Wang, C.: Bridging the Gap in Missing Modalities: Leveraging Knowledge Distillation and Style Matching for Brain Tumor Segmentation . In: proceedings of Medical Image ComputingandComputerAssistedIntervention–MICCAI2025.vol.LNCS15967. Springer Nature Switzerland (September 2025)

work page 2025

[1] [1]

Bai, J., Lekadir, K., Ni, D., Slimani, S., Campello, V.M., Ohene-Botwe, B., Lu, Y., Chen, G., Hou, H., Qiu, D., Zhou, Z.: Intrapartum ultrasound grand challenge 2024 (2024)

work page 2024

[2] [2]

Medical Image Analysis99, 103353 (2025)

Bai, J., Zhou, Z., Ou, Z., Koehler, G., Stock, R., Maier-Hein, K., Elbatel, M., Martí, R., Li, X., Qiu, Y., Gou, P., Chen, G., Zhao, L., Zhang, J., Dai, Y., Wang, F., Silvestre, G., Curran, K., Sun, H., Xu, J., Cai, P., Jiang, L., Lan, L., Ni, D., Zhong, M., Chen, G., Campello, V.M., Lu, Y., Lekadir, K.: PSFHS challenge report: Pubic symphysis and fetal h...

work page 2025

[3] [3]

Medical Image Analysis101, 103420 (2025)

Campello, V.M., Lekadir, K., Bai, J., et al.: Automated angle of progression es- timation from intrapartum ultrasound: A multi-center benchmark study. Medical Image Analysis101, 103420 (2025)

work page 2025

[4] [4]

Scientific Data11(1), 436 (2024)

Chen, G., Bai, J., Ou, Z., Lu, Y., Wang, H.: PSFHS: intrapartum ultrasound image dataset for AI-based segmentation of pubic symphysis and fetal head. Scientific Data11(1), 436 (2024)

work page 2024

[5] [5]

In: Proceedings of the 6th ACM International Conference on Multimedia in Asia

Chen, S., Wang, H., Long, S., Bai, J., Jiang, J.: Ultrasound video segmentation of pubic symphysis and fetal head for AoP measurement. In: Proceedings of the 6th ACM International Conference on Multimedia in Asia. MMAsia ’24, Association for Computing Machinery, New York, NY, USA (2024)

work page 2024

[6] [6]

Computers in Biology and Medicine170, 107917 (2024) 10 Y

Chen, Y., Zhang, C., Chen, B., Huang, Y., Sun, Y., Wang, C., Fu, X., Dai, Y., Qin, F., Peng, Y., Gao, Y.: Accurate leukocyte detection based on deformable-DETR and multi-level feature fusion for aiding diagnosis of blood diseases. Computers in Biology and Medicine170, 107917 (2024) 10 Y. Wang et al

work page 2024

[7] [7]

IEEE Journal of Biomedical and Health Infor- matics28(8), 4648–4659 (2024)

Chen, Z., Lu, Y., Long, S., Campello, V.M., Bai, J., Lekadir, K.: Fetal head and pubic symphysis segmentation in intrapartum ultrasound image using a dual-path boundary-guided residual network. IEEE Journal of Biomedical and Health Infor- matics28(8), 4648–4659 (2024)

work page 2024

[8] [8]

Expert Systems with Applications296, 128998 (2026)

Chen, Z., Ou, Z., Lu, Y., Campello, V.M., Bai, J., Lekadir, K.: Uncertainty- fetal head and pubic symphysis segmentation with enhanced multi-scale features and sparse visual graph attention. Expert Systems with Applications296, 128998 (2026)

work page 2026

[9] [9]

In: Bai, J., Lu, Y

Gan, J., Liang, Z., Fan, J., Mcguire, L., Clarke, J., Cai, W.: Accurate fetal head descent assessment during labor using video Swin Transformer and wavelet-based multitask learning for 2024 MICCAI challenge IUGC. In: Bai, J., Lu, Y. (eds.) Intrapartum Ultrasound. pp. 21–31. Cham (2026)

work page 2024

[10] [10]

In: Linguraru, M.G., Dou, Q., Feragen, A., Giannarou, S., Glocker, B., Lekadir, K., Schnabel, J.A

Jiang, J., Wang, H., Bai, J., Long, S., Chen, S., Campello, V.M., Lekadir, K.: In- trapartum Ultrasound Image Segmentation of Pubic Symphysis and Fetal Head Using Dual Student-Teacher Framework with CNN-ViT Collaborative Learning. In: Linguraru, M.G., Dou, Q., Feragen, A., Giannarou, S., Glocker, B., Lekadir, K., Schnabel, J.A. (eds.) Medical Image Comput...

work page 2024

[11] [11]

Medical Image Analysis96, 103202 (2024)

Jiao, J., Zhou, J., Li, X., Xia, M., Huang, Y., Huang, L., Wang, N., Zhang, X., Zhou, S., Wang, Y., Guo, Y.: USFM: A universal ultrasound foundation model generalized to tasks and organs towards label efficient image analysis. Medical Image Analysis96, 103202 (2024)

work page 2024

[12] [12]

In: Bai, J., Huang, Y., Khobo, I., Yaqub, M., Lekadir, K., Ni, D., Li, S

Liu, X., Hu, J., Li, Y., Chen, X., Wang, Y.: Noisy student-based self-training enhances landmark detection in intrapartum ultrasound. In: Bai, J., Huang, Y., Khobo, I., Yaqub, M., Lekadir, K., Ni, D., Li, S. (eds.) Intrapartum Ultrasound. pp. 1–13. Cham (2026)

work page 2026

[13] [13]

Data in Brief41, 107904 (2022)

Lu, Y., Zhou, M., Zhi, D., Zhou, M., Jiang, X., Qiu, R., Ou, Z., Wang, H., Qiu, D., Zhong, M., Lu, X., Chen, G., Bai, J.: The JNU-IFM dataset for segmenting pubic symphysis-fetal head. Data in Brief41, 107904 (2022)

work page 2022

[14] [14]

Medical Image Analysis109, 103902 (2026)

Ma, X., Tao, Y., et al.: Test-time generative augmentation for medical image seg- mentation. Medical Image Analysis109, 103902 (2026)

work page 2026

[15] [15]

Computers in Biology and Medicine175, 108501 (2024)

Ou, Z., Bai, J., Chen, Z., Lu, Y., Wang, H., Long, S., Chen, G.: RTSeg-net: A lightweight network for real-time segmentation of fetal head and pubic symphysis from intrapartum ultrasound images. Computers in Biology and Medicine175, 108501 (2024)

work page 2024

[16] [16]

IEEE Transactions on Medical Imaging44(6), 2553–2567 (2025)

Ravishankar, H., Paluru, N., Sudhakar, P., Yalavarthy, P.K.: Information Geomet- ric approaches for patient-specific test-time adaptation of deep learning models for semantic segmentation. IEEE Transactions on Medical Imaging44(6), 2553–2567 (2025)

work page 2025

[17] [17]

In: Bai, J., Lu, Y

Tan, Y., Wang, S., Shen, W.D., Yu, W.W., Li, Y., Zhang, P., Jiang, W., Li, Y.J.: Classification and segmentation of intrapartum ultrasound images with deep learn- ing models. In: Bai, J., Lu, Y. (eds.) Intrapartum Ultrasound. pp. 11–20. Cham (2026)

work page 2026

[18] [18]

In: Intrapartum Ultrasound: MICCAI 2025 Grand Challenge, IUGC 2025, Held in Conjunction with MICCAI 2025, Daejeon, South Korea, September 23, 2025, Proceedings

Tang, Y., Zhou, Z., Lu, Y., Bai, J., Long, S., Huang, Y., Khobo, I., Zhang, S., Zhou, Z., Guo, L.: Heatmap regression for automated angle of progression mea- surement: The baseline method for the IUGC2025. In: Intrapartum Ultrasound: MICCAI 2025 Grand Challenge, IUGC 2025, Held in Conjunction with MICCAI 2025, Daejeon, South Korea, September 23, 2025, Pro...

work page 2025

[19] [19]

Medical Physics 52(4), 2311–2324 (2025)

Wang, H., Chen, S., Bai, J.: Temporal modeling of fetal head descent for robust angle of progression estimation in intrapartum ultrasound videos. Medical Physics 52(4), 2311–2324 (2025)

work page 2025

[20] [20]

In: Bai, J., Lu, Y

Xia, Z., Li, H., Lan, L.: DSSAU-Net: U-shaped hybrid network for pubic symphysis and fetal head segmentation. In: Bai, J., Lu, Y. (eds.) Intrapartum Ultrasound. pp. 32–45. Cham (2026)

work page 2026

[21] [21]

Zhou, M., Lu, Y., Bai, J.: Direct regression of the angle of progression in intra- partumultrasoundusingdeepneuralnetworks.ComputersinBiologyandMedicine 173, 108312 (2024)

work page 2024

[22] [22]

Expert Systems with Applications263, 125699 (2025)

Zhou, Z., Lu, Y., Bai, J., Campello, V.M., Feng, F., Lekadir, K.: Segment Anything Model for fetal head-pubic symphysis segmentation in intrapartum ultrasound im- age analysis. Expert Systems with Applications263, 125699 (2025)

work page 2025

[23] [23]

Medical Image Analysis112, 104108 (2026)

Zhu, S., Chen, Y., Chen, W., Jiang, S., Zhou, G., Wang, Y., Qin, F., Wang, C., Tian, Q.: No modality left behind: Adapting to missing modalities via knowledge distillation for brain tumor segmentation. Medical Image Analysis112, 104108 (2026)

work page 2026

[24] [24]

In: proceedings of Medical Image ComputingandComputerAssistedIntervention–MICCAI2025.vol.LNCS15967

Zhu, S., Chen, Y., Chen, W., Wang, Y., Liu, C., Jiang, S., Qin, F., Wang, C.: Bridging the Gap in Missing Modalities: Leveraging Knowledge Distillation and Style Matching for Brain Tumor Segmentation . In: proceedings of Medical Image ComputingandComputerAssistedIntervention–MICCAI2025.vol.LNCS15967. Springer Nature Switzerland (September 2025)

work page 2025