PerkLab / SlicerMatlabBridge

MatlabBridge extension for 3D Slicer
17 stars 8 forks source link

Deformations in the segmentation while reading through nrrdread #6

Open fspinar opened 3 years ago

fspinar commented 3 years ago

Hello there,

I am reading a segmentation (saved as .nrrd file) from Slicer to Matlab by using the nrrdread function (SlicerMatlabBridge/MatlabCommander/commandserver/nrrdread.m)

However, the segments are not entirely correct after I read them. It seems like their boundaries are not read precisely, i.e., the middle or interior parts are correct, but as I go through the boundaries there are some errors.

Any suggestions ? I am using Slicer 4.11 and Matlab 2020b.

Thanks !

lassoan commented 3 years ago

There should be no such errors. Please provide more informatíon (screenshots, sample data, etc).

Note that we added MatlabBridge extension to Slicer about 5-10 years ago. It made sense then. Python is so much more capable nowadays that I would strongly recommend you to switch to Python as soon as you can.

fspinar commented 3 years ago

Screenshot_8

This is one example screenshoot. Two fiducial points which should be inside of the orange segment are not assigned to that segment. Also when I "manually" checked their MRI coordinates in Matlab (by using the given nrrdread function), it seems like that MRI coordinate does not correspond to that segment. So must to be about the nrrdread function itself, not the way I use it in my scripts.

This usually happens around the boundaries of the segments, but here it also happens in a point which is quite inside of the segment.

Any suggestions ?