Published October 25, 2023 | Version v4
Dataset Open

MRI Neonatal Lung Segmentation and 3D Morphologic Features

Authors/Creators

Description

We developed an ensemble of deep convolutional neural networks (2D-UNets) to perform automated neonatal lung segmentation from MRI sequences. A three-dimensional reconstruction is used to calculate MRI features for lung volume, shape, pixel intensity, and surface.

In addition, ML Models for severity prediction of Bronchopulmonary Dysplasia (BPD) are implemented as an applied example of the use of MRI lung volumetric features for disease prognosis.

This dataset comprises:

  • Three pretrained 2D-UNet Models for Neonatal MRI Lung Segmentation.
  • Resulting performances and features per MRI-sequence.

See Publication:

Automated MRI Lung Segmentation and 3D Morphologic Features for Quantification of Neonatal Lung Disease (2023)

https://doi.org/10.1148/ryai.220239

Files

Files (373.5 MB)

Name Size Download all
md5:e9b510af1f13bb4a303fd94ff32a0726
141.0 kB Download
md5:866c5200246114d741ee3a6cdbbcb926
111.9 kB Download
md5:eb1075ac553379abaf2553fd1ca8f86e
26.3 kB Download
md5:22127262818b7cde2d68bc85ebe82ebf
124.4 MB Download
md5:67d0444e0b951c4d11dec2957a700c15
124.4 MB Download
md5:3db1edef9739574afb6b155600205ff2
124.4 MB Download

Additional details

Related works

Is published in
Publication: 10.1148/ryai.220239 (DOI)

Software

Repository URL
https://github.com/SchubertLab/NeoLUNet

References