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
1.18k stars 498 forks source link

2d_wavelet_features_extract #398

Closed Bercy0616 closed 6 years ago

Bercy0616 commented 6 years ago

Excuse me I added "force2d:true and force2Ddimension: 0" in params When I extracted the features, the features name is [HH.HL,LH,LLL] LLL????? = =?

I wanna know if the LLL is LL?

paste my code: params = '...\exampleCT.yaml' extractor = featureextractor.RadiomicsFeaturesExtractor(params) extractor.enableImageTypes(Wavelet={})

result = extractor.execute(ori_file_path, seg_file_path)

fedorov commented 6 years ago

Looks like a bug, but I didn't manage to track it down yet... Thank you for your report!

JoostJM commented 6 years ago

LLL in this case is indeed LL, why the name is not rendered correctly I do not know, but I'll look into it.

BalthasarSchachtner commented 6 years ago

Hi, the "LLL" is coming from the fact, that the name of the "approximation" is hard coded here: https://github.com/Radiomics/pyradiomics/blob/2a9fd79520cfc1330f714112bd1eaf7dd63e7ec7/radiomics/imageoperations.py#L706 I came across this issue today since I was wondering why the LL/LLL wavelet decomposition is handled separately? As far as I can see it will be used in the same way as the other combinations of high- and low-pass filters LH/HL/HH (and their 3D companions).

JoostJM commented 6 years ago

@BalthasarSchachtner, this has to do with the fact that when you do multi-level wavelet decomposition, generally only the last approximation is returned.