Open tangy5 opened 1 year ago
This is a great initiative, @tangy5 For robustness, let's use two-stage approaches in which we first segment the organ and then focus on the organ tumours/vasculature. For organ segmentation, we can use the TotalSegmentator dataset. Happy to help!
Hi @diazandr3s
On the AIAA 3D Slicer annotation server, there's a liver and tumor CT segmentation model available, I'm wondering why this model is not in the MONAI zoo
Here's a link to the mentioned model :
https://catalog.ngc.nvidia.com/orgs/nvidia/teams/med/models/clara_ct_seg_liver_and_tumor_amp
Maybe that's what you are looking for (a liver tumor foundation model) @tangy5
We already have two foundation models for basic whole brain and whole body segmentation.
It would be great to add tumor segmentation foundation model for liver,pancreas,lung,colon,hepatic vessel, etc. Use MSD datasets for training, the goal is to train a single to output many types of tumor, which can be used as a base model for tumor detect and seg tasks.
Users can use the foundation model for fine-tuning specific tumor/nodule in practice workflow.