How to threshold only one slice in the DICOM Module

Hi all,

I have loaded a 3D- CT scan into Slicer using the DICOM module. How can I apply thresholding to only one slice?

Thanks

There are many ways to do it. What would you like to achieve?

Hi Andras,

Thanks for the reply. I want to achieve accurate segmentation of muscle, SAT and VAT of the entire thoracolumbar region from T1 - L5. The method I was planning was to threshold 17 mid-vertebra slices and manually correcting them, then using the Fill between slice effect to propagate to the other slices and finally manually correcting them again. Is this a solid method to achieve the highest accuracy segmentation for a high resolution CT-volume? I am open to other suggestions :).

Thanks

Manual correction typically takes much longer time than specifying a region from scratch. So, I would definitely not recommend to create an initial “segmentation” with simple thresholding.

Instead, I would recommend to try these options:

  • Paint effect with appropriate “Editable intensity range”. You can find the correct threshold range in Threshold effect and then click “Use for masking”.
  • Draw effect, potentially with “Editable intensity range”.
  • Level tracing effect.

You may try to pre-process the image to reduce noise and make regions more homogeneous. For example, try Anisotropic Diffusion filters (in Filtering / Denoising category), or various other filters in Simple Filters module.

Great thanks for the suggestions! I still want to experiment thresholding at a single slice. You mentioned there are many ways to do this for a CT scan loaded using the DICOM module?

The easiest to see how bad the result is if you use simple thresholding is to threshold the entire volume, using Threshold effect. If you really want to threshold only a handful of slices, then you can set editable intensity range in Threshold effect, click Use for masking, then use Paint effect with a huge brush (so that you can easily cover the entire slice with a single click or a short paintstroke).

1 Like

Hi Andras,

Your fast responses are always much appreciated! Thanks.

1 Like