Since MedSAM is just SAM with different pretrained weights, the tool backend can share the same virtual environment.
It would also conceptually make sense to reuse the same python wrapper daemon created for SAM to preprocess images from MITK, for MedSAM as well.
See "agent sam" here: https://github.com/MIC-DKFZ/agent-sam
Description
Description
| Status | Assigned | Task | ||
|---|---|---|---|---|
| Open | None | T29191 [SEG] New segmentation tool candidates | ||
| Open | None | T30314 MedSAM in MITK | ||
| Open | a178n | T30354 [MedSAM] Update SAM python daemon script for MedSAM |