This page contains submission procedure for the segmentation results. For paper submission guidelines to the STACOM 2011 workshop, please refer to the website.
Participants must send their results as binary images (similar to the ground truth images), where the value of each pixel is one of the following:
0 = background (non-object) structure
non zero = myocardium
The size of the image matrix must be identical with the input image slice. Following clinical convention, papillary muscles are excluded from the myocardium. Only segmentation results from the validation set are needed. Participants have full flexibility to choose whether to use both short-axis and long-axis, or only one of them.
In addition to the segmentation images, participants should also send LV volumes and masses information in order to assess the global LV functions which are going to be validated with the ground truth volume and mass values. Without this information, we cannot assess the clinical measurements.
Create segmentation images in NRRD (Nearly Raw Raster Data) or PNG formats with similar filename with the input image filename, prefixed with 3 characters to identify the participant ID.
Create a separate CSV (comma separated values) file, where each row contains the following 6 values: caseID,EDV,phED,ESV,phES,M where:
caseID is the data set name, that is the folder name where the data is located in the validation set,
EDV = End-Diastolic endocardial volume in ml,
phED is an integer between 1 to n where EDV was calculated at (most likely phase 1),
ESV = End-Systolic endocardial volume in ml,
phES is an integer between 1 to n where ESV was calculated at, and