Handling of mismatched size mask is still not working

[fedorov]

I think the issue is that the resampled mask does not get propagated from here: https://github.com/Radiomics/pyradiomics/blob/master/radiomics/imageoperations.py#L238

checkMask returns the bounding box for the resampled mask here: https://github.com/Radiomics/pyradiomics/blob/master/radiomics/featureextractor.py#L326

But cropToTumorMask is operating on the original mask before resampling: https://github.com/Radiomics/pyradiomics/blob/master/radiomics/featureextractor.py#L340

Identified by Dev Gude.

[fedorov]

@JoostJM I suggest to fix this by having checkMask return a tuple of bounding box and an updated mask, if it had to be resampled, and then use that updated mask after the mask check. Do you have objections?

[JoostJM]

No objections, saves me the trouble ;-)

…

-------- Oorspronkelijk bericht -------- Van: Andrey Fedorov <[email protected]> Datum: 25-05-17 20:09 (GMT+01:00) Aan: Radiomics/pyradiomics <[email protected]> Cc: Joost van Griethuysen <[email protected]>, Mention <[email protected]> Onderwerp: Re: [Radiomics/pyradiomics] Handling of mismatched size mask is still not working (#259) @JoostJM<https://github.com/joostjm> I suggest to fix this by having checkMask return a tuple of bounding box and an updated mask, if it had to be resampled, and then use that updated mask after the mask check. Do you have objections? - You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#259 (comment)>, or mute the thread<https://github.com/notifications/unsubscribe-auth/ARMRw_K75SeakVoZEMe8xLxZrA-cC5gwks5r9cPugaJpZM4NlWkS>.