A multicentre benchmark dataset for comprehensive landmark-based fetal ultrasound biometry
Chiara Di Vece and Zhehua Mao and Netanell Avisdris and Brian Dromey and Raffaele Napolitano and Dafna Ben Bashat and Francisco Vasconcelos and Danail Stoyanov and Leo Joskowicz and Sophia Bano

@article{figshare_30819911,
  title= {A multicentre benchmark dataset for comprehensive landmark-based fetal ultrasound biometry},
  journal= {},
  author= {Chiara Di Vece and Zhehua Mao and Netanell Avisdris and Brian Dromey and Raffaele Napolitano and Dafna Ben Bashat and Francisco Vasconcelos and Danail Stoyanov and Leo Joskowicz and Sophia Bano},
  year= {2026},
  url= {https://rdr.ucl.ac.uk/articles/dataset/A_multi-centre_multi-device_benchmark_dataset_for_landmark-based_comprehensive_fetal_biometry/30819911},
  abstract= {Multicentre-Fetal-Biometry is a comprehensive benchmark dataset for landmark-based fetal biometry estimation from 2D ultrasound, comprising 4,513 de-identified ultrasound images from 1,904 subjects acquired at four clinical sites using seven different ultrasound devices. The dataset provides expert anatomical landmark annotations for all clinically used fetal biometric measurements: head bi-parietal diameter (BPD), occipito-frontal diameter (OFD), transverse abdominal diameter (TAD), anterior-posterior abdominal diameter (APAD), and femur length (FL). The dataset is derived from three existing sources: the Fetal Plane (FP) dataset, the HC18 head dataset, and a dataset acquired at University College London Hospital (UCLH). All images are standard 2D ultrasound planes captured by experienced sonographers following ISUOG guidelines. The anatomical landmarks were produced by expert clinicians in the source datasets — manual landmark annotation for the FP and UCL subsets, and landmarks derived via ellipse fitting from segmentation masks for the HC18 subset — and we harmonised these into a single, consistent landmark format for release. For HC18, image-centric preprocessing parameters were recomputed to ensure that all landmarks lie within the model's effective field of view, correcting an earlier landmark-space mismatch. For each image, we provide precise landmark coordinates that define the start and end points of each biometric measurement. Each sample in FetalBiometry-MultiCentre-Landmarks consists of: A 2D ultrasound image Landmark coordinates for fetal head measurements (BPD, OFD) when applicable Landmark coordinates for fetal abdomen measurements (TAD, APAD) when applicable Landmark coordinates for femur measurement (FL) when applicable Metadata including: pixel-to-millimetre conversion rate, de-identified subject ID, anatomy (head, abdomen, or femur), device/acquisition identifier where available, and subject-disjoint train/test split assignment. The dataset emphasizes challenging but clinically relevant cases, including: Multicentre variability with different ultrasound devices (GE Voluson E6, S8, S10, E8, 730; Aloka) Operator-dependent variability in probe handling and image framing Fetal presentation variability across gestational ages Anatomical variability in landmark positions, sizes, and orientations Cross-domain differences between manual landmark annotation (FP, UCL) and ellipse-derived landmarks (HC18) Multicentre-Fetal-Biometry is designed to support rigorous and reproducible benchmarking of automated fetal biometry algorithms, particularly for evaluating domain shift and cross-centre generalisation in image-guided obstetric ultrasound. Citation If you use this dataset, please cite: Di Vece, C., Mao, Z., Avisdris, N. et al. A multicentre benchmark dataset for comprehensive landmark-based fetal ultrasound biometry. Sci Rep 16 , 17405 (2026). https://doi.org/10.1038/s41598-026-47854-3 Declarations Our dataset is released under the CC BY-NC-SA 4.0 license and may only be used for non-commercial purposes.

**License:** https://creativecommons.org/licenses/by-nc-sa/4.0/. **Source:** https://rdr.ucl.ac.uk/articles/dataset/A_multi-centre_multi-device_benchmark_dataset_for_landmark-based_comprehensive_fetal_biometry/30819911

**Related papers:**
- Di Vece, C., Mao, Z., Avisdris, N., et al. A multicentre benchmark dataset for comprehensive landmark-based fetal ultrasound biometry. Scientific Reports 16, 17405 (2026). arXiv:2512.16710 [doi:10.1038/s41598-026-47854-3](https://doi.org/10.1038/s41598-026-47854-3) · [Semantic Scholar](https://www.semanticscholar.org/paper/df76e17d2007027307447d763f886e8a7a53da73)},
  files= {Multicentre-Fetal-Biometry-2026.zip},
  keywords= {Fetal Ultrasound Imaging, Fetal biometry, Domain Shift, Artificial intelligence (AI) models},
  license= {https://creativecommons.org/licenses/by-nc-sa/4.0/},
}