Published June 4, 2020 | Version v1
Dataset Open

A Deep Learning Based Cardiac Cine Segmentation Framework for Clinicians - Transfer Learning Application to 7T - Additional Data

  • 1. Chair of Cellular and Molecular Imaging, Comprehensive Heart Failure Center, University Hospital Würzburg

Description

Automatically Generated Segmentation Masks for Data Science Bowl Cardiac Challenge Data

These segmentation masks have been automatically generated with the ukbb_cardiac network by Bai et al. (2018, doi:10.1186/s12968-018-0471-x). In order to reproduce the data cleaning and conversion: download the data from Kaggle and follow the data curation and conversion steps outlined in: cmr-seg-tl and the associated publication (Link to be added).

As this is a derived dataset please abide by the data use rules of the original dataset at kaggle and provide proper citation to the original data:

The data for the Data Science Bowl is available for research and academic pursuits. Please cite as ‘Data Science Bowl Cardiac Challenge Data’.

Please also cite the Bai et al. article for the algorithm and our publication for the data curation.

Notes

Please be aware that the masks were automatically generated, not quality controlled. Validation has been performed against ground truth volume information from kaggle. This can be used as a measure of confidence but does not replace proper quality control.

Files

data.zip

Files (19.3 GB)

Name Size Download all
md5:00f0d8cfe51a1fb8ed489ebf2161e06f
19.3 GB Preview Download

Additional details

Related works

Cites
Journal article: 10.1186/s12968-018-0471-x (DOI)