fidoptlungmask
lungmask's U-Net(R231), trained on 231 routine CTs of varied disease, the right and left lungs slice by slice, run as its authors' command line runs it.
SegmentationGPUversion lungmask 0.2.21, weights unet_r231-d5d2fc3d.pth and unet_ltrclobes-3a07043d.pth of release v0.0 (SHA-256 checked)lung, thorax, thorax, abdomen, thorax, abdomen and pelvis, whole body · CTApache-2.0
Run it
With its KonfAI app, at the benchmark's settings (the fidopt app repositories are private for now: a Hugging Face token with access opens them):
pip install konfai-apps
| Configuration | Command | |
|---|---|---|
| lungmask | Get the app | konfai-apps infer fideus-labs/fidopt-segmentation:Lungmask -i image.nii.gz -o out --gpu 0 |
| lobes | Get the app | konfai-apps infer fideus-labs/fidopt-segmentation:Lungmask_Lobes -i image.nii.gz -o out --gpu 0 |
Where it works
Its place on each kind of problem it has results on, among the methods ranked there. A method that leads one problem can trail another: read the one closest to yours.
| Problem | Place | Ranked | Datasets |
|---|---|---|---|
| Body CT, many structures, segmentation | not ranked yet | 3 of 5 | |
| Body CT, many structures, segmentation | not ranked yet | 1 of 5 | |
| Chest CT, segmentation | not ranked yet | 1 of 4 | |
| Chest CT, segmentation | not ranked yet | 1 of 4 |
Results
| Dataset | # | Read by | Score | Failed | Time s |
|---|---|---|---|---|---|
| COVID-19 CT Lung and Infection | 3 | Dice | 0.571 | 0 | 160 |
| COVID-19 CT Lung and Infection | 5 | Dice | 0.571 | 0 | 42 |
| CT-ORG | 1 | Dice | 0.975 | 0 | 281 |
| LCTSC, thoracic organs at risk | 1 | Dice | 0.944 | 0 | 14 |
| LCTSC, thoracic organs at risk | 2 | Dice | 0.931 | 0 | 63 |
| Pediatric-CT-SEG, organs of children | 1 | Dice | 0.922 | 0 | 20 |
| Pediatric-CT-SEG, organs of children | 2 | Dice | 0.896 | 0 | 118 |
| TotalSegmentator CT | 4 | Dice | 0.657 | 0 | 48 |
- Reference
- Hofmanninger et al., European Radiology Experimental 2020