Low resolution when importing an STL segmentation

Support
, , , , , ,
1

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

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

Here are a few images:

Before import
Before import4264×2568 386 KB

After conversion
After conversion1920×1189 153 KB

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!

2

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.