You need to agree to share your contact information to access this dataset

This repository is publicly accessible, but you have to accept the conditions to access its files and content.

This dataset contains de-identified maternal-fetal ultrasound images released under CC BY 4.0. Access is granted for research and educational use. By requesting access you agree to cite the original authors and not to attempt re-identification of any subject.

Log in or Sign Up to review the conditions and access this dataset content.

Fetal Planes DB

A mirror of FETAL_PLANES_DB: 12,400 routinely-acquired maternal-fetal screening ultrasound images from two hospitals, captured by several operators on several machines and manually labelled by an expert maternal-fetal clinician.

Contents

Field Description
image The ultrasound image
Plane Abdomen, Brain, Femur, Thorax, Maternal cervix, Other
Brain_plane Trans-thalamic, Trans-cerebellum, Trans-ventricular, Other, Not A Brain
Patient_num Subject id — group by this when splitting
Train 1 = train, 0 = test (the authors' official, patient-disjoint split)
US_Machine, Operator Acquisition metadata

Splits follow the original Train column.

Usage

from datasets import load_dataset
ds = load_dataset("shr3m/fetal-planes-db")   # requires access approval

Notes

  • Classes are imbalanced; Other is a large catch-all that the FetalCLIP zero-shot benchmark excludes.
  • Images vary in size and are mostly non-square — pad to square before resizing, do not stretch.
  • The official split is patient-disjoint. Preserve that if you re-split.

Licence and citation

CC BY 4.0. Attribution is required. This is a mirror; all credit belongs to the original authors.

Original dataset: DOI 10.5281/zenodo.3904280

@article{burgos2020evaluation,
  title   = {Evaluation of deep convolutional neural networks for automatic
             classification of common maternal fetal ultrasound planes},
  author  = {Burgos-Artizzu, Xavier P. and Coronado-Guti{\'e}rrez, David and
             Valenzuela-Alcaraz, Brenda and Bonet-Carne, Elisenda and
             Eixarch, Elisenda and Crispi, Fatima and Gratac{\'o}s, Eduard},
  journal = {Scientific Reports},
  volume  = {10},
  number  = {1},
  pages   = {10200},
  year    = {2020},
  doi     = {10.1038/s41598-020-67076-5}
}
Downloads last month
11