I am curious if there is a module to match segmentations to input volume images in order to train a neural network?
I have converted DICOM to PNG and converted segmented images into labelmaps and numpy arrays. The arrangement of segmented masks are not in order of the original image data.
That’s not a good operation because you loose the spatial reference system of the data. MonaiLabel extension in slicer allows you to train a segmentation model by simply providing the 3D images and the associated segmentations, all in Slicer format. No need to convert anything to anything.
Thank you for your response and sorry for the late reply.
I am not sure I am allowed to use the Preview Release version of 3D slicer for my project.
Is there a way to work around this without using Monai?