search

rbumm / SlicerLungCTAnalyzer

This is a 3D Slicer extension for segmentation and spatial reconstruction of infiltrated, collapsed, and emphysematous areas in lung CT.
Apache License 2.0
71 stars 23 forks source link

Which kernel do we use ? #21

Closed rbumm closed 3 years ago

rbumm commented 3 years ago

Writing an abstract my radiologist coauthor asked, which kernel we use for our calculations. @lassoan  Will gladly name you as a coathor on following abstracts and papers. Can the we specify, which kernel we used in CTLA or can we change the kernel if needed? Thank you

lassoan commented 3 years ago

I'm not sure what kernel you mean. For which operation?

rbumm commented 3 years ago

My radiologist spoke about the fact that her emphysema judgement was "affected by the type of kernel" she used. 

In 3D slicer, I saw the possibility to select a "kernel" in the chest imaging platform under -> airway segmentation. Concerning Lung CT Analyzer - I am not sure how to answer her.

lassoan commented 3 years ago

Probably by "kernel" your radiologist means which interpolation or additional filtering method is used.

By default, Slicer uses bilinear interpolation for computing displayed pixels from voxels of the CT volume (with a few lines of Python script we can also switch to bicubic or sinc interpolation). Additional image processing filters are available in Simple Filters and other modules.

If your radiologist does not like the simple bilinear interpolation display and has some time to look at various interpolation options then I can provide Python code snippets that can switch between the different options. Note that resolution and color reproduction capabilities of the display may matter, too, so most likely a random laptop will not be able to show as nice images as displays in a radiologist reading room.