search

MHubAI / models

Stores the MHub models dockerfiles and scripts.
MIT License
8 stars 16 forks source link

BAMF FDG-Avid Breast Tumor #79

Closed jithenece closed 3 months ago

vanossj commented 5 months ago

@LennyN95 What is the right way to have multi modality inputs into the NNUnetRunner? We made a custom runner, but it doesn't feel like that should be required.

LennyN95 commented 5 months ago

@vanossj if it's an nnunet that takes e.g. multiple modalities interchangeably but only one input at a time, then you could use the | operator like in nifti:mod=mr|ct. However, if you have multiple inputs (as your implementation suggests), the current version of the NNUnet runner indeed isn't capable of handling this. We are planning to update this (however, on a lower prior). So, in this case, a custom module (you can, of course, start from our implementation and simply make the desired modifications) would be the fastest.

vanossj commented 5 months ago

@LennyN95 Could you add a roi for FDG-avid tumor? (NCIt C168968). if you have an alternative sct code that would work too

LennyN95 commented 5 months ago

@jithenece @vanossj As per your suggestion, I added FDG_AVID_TUMOR to SegDB.

jithenece commented 4 months ago

@LennyN95 could you add BREAST to segdb as it is required for generating roi=BREAST+FDG_AVID_TUMOR.

LennyN95 commented 4 months ago

@jithenece done, added BREAST as new segmentation to SegDB.

jithenece commented 4 months ago

I tried to test this but seems to fail. Does this require updating base image mhubai/base:latest. sha256:d72b350c7f98e5f2460297d6059a70f967091c9ed11728ab3bbc7f9b32b65d2e

@jithenece done, added BREAST as new segmentation to SegDB.

jithenece commented 3 months ago

closing this as this will be done in another pull request