| --- |
| license: other |
| license_name: vesselfm-open-rail-plus-plus-m |
| license_link: LICENSE |
| tags: |
| - medical-imaging |
| - segmentation |
| - 3d |
| --- |
| |
| # vesselfuse |
|
|
| Weights for [`vesselfuse`](https://github.com/eigenvivek/vesselfuse): multimodal super-resolved vessel segmentation. |
|
|
| - `vesselfuse.pt`: PyTorch state dict for `vesselfuse.VesselFuse` (MONAI DynUNet, 18 input channels: 9 images and 9 presence masks for CTA, CT, TOF, T1, T1+c, T2, FLAIR, SWI, SWI phase). |
|
|
| The `vesselfuse` CLI downloads these weights on first use. |
|
|
| ## License |
|
|
| The weights are fine-tuned from the [vesselFM](https://huggingface.co/bwittmann/vesselFM) weights and are distributed under the vesselFM Open RAIL++-M License ([LICENSE](LICENSE)). |
| Use is restricted to research and non-commercial purposes (Attachment A). |
|
|
