Jan 13, 2021
Do you also wonder what the effect is of differences in segmentations and segmentation programs on the inverse problem of electrocardiography? Within the Model Building working group of the Consortium for ECG imaging, we try to answer this question and with that we need your help!
Four imaging datasets were selected and thereof already several people segmented the heart (endocardium and epicardium), lungs and thorax, wherefore we are very grateful! However, more segmentations are required and therefore we want to ask for your help.
Each dataset contains its own challenge, like imaging artifacts or segmentation without contrast. On the Google Drive, the four datasets are uploaded as nrrd and DICOM files:
1. Dalhousie: segmentation of blood pools, epicardium/endocardium, lungs and thorax.
2. P43: segmentation of blood pools, epicardium/endocardium, lungs and thorax, challenge with imaging artifacts.
3. Nijmegen: segmentation of blood pools, epicardium/endocardium, lungs and thorax
4. Auckland: segmentation of epicardium, lungs and thorax, challenge with imaging artifacts and without blood pool contrast.
If you are in the opportunity to segment one or more of these datasets using the images uploaded at the drive with your own segmentation tools before the 10th of February, it would really help us making another step forward in the analysis towards answering our question! Thanks in advance for your help and for the segmentations, we really appreciate it! If you have any questions, don’t hesitate to ask me (Machteld, m.j.boo...@umcutrecht.nl)!