Background

During pregnancy, ultrasound imaging is used to measure fetal biometrics. One of these measurements is the fetal head circumference (HC). The HC can be used to estimate the gestational age and monitor growth of the fetus.

The HC is measured in a specific cross section of the fetal head, which is called the standard plane. The dataset for this challenge contains a total of 1,334 two-dimensional (2D) ultrasound images of the standard plane that can be used to measure the HC.

This challenge makes it possible to compare developed algorithms for automated measurement of fetal head circumference in 2D ultrasound images.

Detailed Description of the Data

The data for this challenge can be downloaded from Zenodo: DOI 10.5281/zenodo.1322001

• The data is divided into:
• Training set: 999 images

• Test set: 335 images

• Each 2D ultrasound image is 800 × 540 pixels, with a pixel size ranging from 0.052 to 0.326 mm.

• Pixel sizes for each image are listed in:
• training_set_pixel_size_and_HC.csv

• test_set_pixel_size.csv

• The training set includes an image with manual annotation of the HC made by a trained sonographer. The corresponding measurement (in mm) is also included in training_set_pixel_size_and_HC.csv.

• Note: All filenames start with a number. Although there are 999 training images, the filenames only go up to 805. Some images were captured during the same examination and appear very similar. These have an additional number in the filename between _ and HC, for example:

• 010_HC.png
• 010_2HC.png

Challenge

This challenge is aimed at designing an algorithm that can automatically measure the fetal head circumference given a 2D ultrasound image.

You can use the 999 ultrasound images and annotations in the training set to develop your algorithm. The final model should be applied to the independent test set of 335 images.

Results can be uploaded under the Submit tab. They will be evaluated automatically and displayed on the leaderboard under the Results tab, allowing you to compare your performance with other submitted algorithms.

Cite As

When referencing this challenge, please include the following citations:

• Thomas L. A. van den Heuvel, Dagmar de Bruijn, Chris L. de Korte and Bram van Ginneken. Automated measurement of fetal head circumference using 2D ultrasound images. PLOS ONE, 13(8), 2018: e0200412.
• Thomas L. A. van den Heuvel, Dagmar de Bruijn, Chris L. de Korte and Bram van Ginneken. Automated measurement of fetal head circumference using 2D ultrasound images [Data set]. Zenodo. https://doi.org/10.5281/zenodo.1322001

Submission Instructions

Submit your results as a CSV file with:
- 6 columns
- 336 rows (1 header row + 335 test samples)

CSV Header:

filename,center_x_mm,center_y_mm,semi_axes_a_mm,semi_axes_b_mm,angle_rad

Each row describes the ellipse for one test image:

• filename: must include .png (e.g., 001_HC.png)
• center_x_mm, center_y_mm: center of the ellipse (in mm)
• semi_axes_a_mm, semi_axes_b_mm: lengths of the semi-axes (in mm)
• angle_rad: orientation of the ellipse (in radians)

Example (from 388_HC.png):

388_HC.png,72.378073183,60.803647248,32.482252220,23.374422688,0.061252332740