Low resolution when importing an STL segmentation

[ludoludo (Ludovic Corcos)]

Hey!

I have a Nifti volume with the following dimensions: 588 x 734 x 1095, with an isometric resolution of 0.033 mm.

I’ve got an STL (or OBJ) file that represents a segmentation I’ve already made.

I’d like to convert this segmentation into a binary map that I can edit at high resolution: image spacing ~ 0.01.

The problem is that when I import this file and convert it directly, I end up with a very pixelated mask even at this resolution.

I’ve tried different approaches as indicated on the forum:
https://slicer.readthedocs.io/en/latest/user_guide/modules/segmenteditor.html#segmentation-is-not-accurate-enough

Segmentation resolution

Hello Can I select the pixel resolution of my segmentation? is it the same as native image resolution? thanks

How to increase the segmentation resolution?

Hi everybody, How can I increase the resolution of the segmentation as the default one uses large size of voxels (cubes)? Is this called oversampling? how to do it without changing the volume as I need to calculate the volume afterwards? Thanks,

Increasing segmentation resolution

Hello all, How can I increase the resolution of a current segmentation? The master volume has its own quality and we cannot increase it but I want a segmentation with higher quality. I tried to crop the master volume with the spacing scale of 0.5 and with both linear and B-spline interpolators. However, it seems not to be working. Is this the correct way of doing what I am looking for or there other ways too? Thanks in advance.

I tested several versions by increasing the dimensions of the STL/OBJ file, but it doesn’t change anything

Here are a few images:

The idea is to have a high-resolution 3D segmentation mask that I can register on my microCT so that it matches perfectly.

Thanks for your help!

[lassoan (Andras Lasso)]

See step-by-step instructions here.

Note that an image of 588 x 734 x 1095 at 0.033mm/pixel resampled to 0.01mm/pixel would result in a 12GB image. For processing a segmentation of this size it would be recommended to have about 100GB physical RAM for optimal performance.