Dataset for Intrapartum Ultrasound Grand Challenge (IUGC 2025) in conjunction with MICCAI 2025

In 2018, the World Health Organization (WHO) issued 56 recommendations aimed at improving the quality of intrapartum care and enhancing women’s childbirth experiences. Building on these recommendations, the WHO introduced the Labour Care Guide (LCG) in 2020—a next-generation tool designed to support evidence-based, respectful, and woman-centered care during labor and delivery. Developed through expert consultations, field research, and usability studies across multiple countries, the LCG provides a structured framework for monitoring labor progress and maternal–fetal well-being by recording key clinical parameters. When deviations from normal labor progression occur, the LCG highlights these abnormalities to prompt timely and appropriate clinical interventions. Intrapartum ultrasound plays a critical role in this context, providing quantitative and reproducible assessment of labor progress. The core operation of such ultrasound-based assessment lies in the accurate identification of anatomical landmarks, which form the foundation for subsequent measurements of angles and distances—most notably the Angle of Progression (AoP) and Head–Symphysis Distance (HSD). These quantitative parameters are vital for diagnosing labor arrest and guiding decisions on intervention timing and delivery mode. However, manual annotation of these landmarks is highly dependent on expert experience, time-consuming, and subject to inter-operator variability. As such, there is a pressing need for fully automated and precise landmark localization methods to improve efficiency, reproducibility, and clinical adoption. To address this challenge, the Intrapartum Ultrasound Grand Challenge (IUGC) 2025 was established as a collaborative initiative involving the Deep Learning in Intrapartum Ultrasound Image Analysis consortium and leading clinical organizations, including the International Society of Ultrasound in Obstetrics and Gynecology (ISUOG), the World Association of Perinatal Medicine (WAPM), the Perinatal Medicine Foundation (PMF), and the National Institute for Health and Care Excellence (NICE). The overarching goal is to design clinically relevant benchmarks that foster the translation of cutting-edge algorithms into real-world obstetric practice. Since its inception at MICCAI 2023, the IUGC series has progressively expanded in scope and complexity. The 2023 edition focused on the Pubic Symphysis–Fetal Head Segmentation (PSFHS) task, providing a well-annotated dataset for segmentation-based biometry. In MICCAI 2024, the challenge evolved into a multi-task benchmarking platform by: (1) extending analysis from static images to dynamic videos; (2) expanding tasks from segmentation to classification, biometry, and measurement; (3) incorporating multiple quantitative parameters (AoP and HSD); and (4) broadening data sources from Asia to include Europe and Africa. This evolution has established a robust benchmarking ecosystem enabling systematic comparison of algorithms across diverse tasks, modalities, and clinical settings. IUGC 2025 represents the next major step in this progression. It focuses explicitly on end-to-end landmark-based biometry for comprehensive assessment of labor progress. Key innovations include: Task: A semi-supervised landmark detection framework that eliminates reliance on segmentation outputs. Dataset: A large-scale, multi-center dataset covering all fetal descent stations (five “minus,” one “zero,” and three “plus” levels) from more than 20 clinical institutions, comprising 28,919 ultrasound images and videos. The training set includes 300 labeled and 31,421 unlabeled cases, with 100 validation and 501 hidden test cases. Evaluation: Metrics emphasize detection accuracy (Mean Radial Error), measurement precision (absolute parameter deviation), and inference efficiency. A subset of test cases is independently annotated by multiple raters to compare algorithmic performance with human inter-rater variability. Through its scale, design, and openness, IUGC 2025 establishes a new benchmark for intelligent labor monitoring—bridging the gap between algorithmic development and clinical implementation, and paving the way toward automated, objective, and reproducible intrapartum ultrasound assessment.