search

AIM-Harvard / pyradiomics

Open-source python package for the extraction of Radiomics features from 2D and 3D images and binary masks. Support: https://discourse.slicer.org/c/community/radiomics
http://pyradiomics.readthedocs.io/
BSD 3-Clause "New" or "Revised" License
1.14k stars 492 forks source link

Feature extraction for bin-width 255 #727

Open sovanlal opened 2 years ago

sovanlal commented 2 years ago

Hi,

I was extracting radiomics features from a CT dataset at various bin-widths for a comparison study (25, 32, 255 etc).

I was comparing individual gray level features between bin-width 25 and 255. My CT images are normalized to intensity levels (0 to 255). So, for bin-width 255, I just have 1 gray level.

While comparing GLCM between bin-width 25 and 255 , I noticed that GLCM for bin-width 255 is almost non-existing which makes sense since I have only one gray level. However, while comparing GLRLM between bin-width 25 and 255, I see a similar pattern between bin-width 25 and 255 although the calculated feature-values are different. I have attached three slides describing what is happening.

My question is for bin-width 255, both GLCM and GLRLM should behave in a similar way since we are dealing with only one gray level. But why, GLRLM feature pattern at bin-width 255 are similar to bin-width 25?

Could you please help me figure it out?

Version (please complete the following information):

Thank you,

JoostJM commented 2 years ago

I don't see any attachments, so I can't be sure. Can you review the log on how many gray values are computed for binwidth 25? Can you share example data that is giving you this error (and the script you use to run this)?