MONAI wholeBody CT

MONAI's wholeBody_ct_segmentation bundle, a SegResNet trained on the TotalSegmentator dataset at 1.5 mm: TotalSegmentator v1's 104 structures of a CT, run as the bundle's inference.json runs it.

SegmentationGPUversion bundle 0.2.7, models/model.pt (sha256 80b429fb) and model_lowres.pt (sha256 c3ab55eb), MONAI 1.5.1CTApache-2.0

Get the app Source

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
MONAI wholeBody CT	Get the app	`konfai-apps infer fideus-labs/fidopt-segmentation:MONAI_WholeBody -i image.nii.gz -o out --gpu 0`
3 mm	Get the app	`konfai-apps infer fideus-labs/fidopt-segmentation:MONAI_WholeBody -i image.nii.gz -o out --gpu 0 --set=highres=false`

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
Abdomen CT, segmentation	3	of 4	12 of 12
Liver segmentation on CT without physical geometry	5	of 5	1 of 1
Clinical pelvic bones, CTPelvic1K	5	of 5	1 of 1
Body CT, many structures, segmentation		not ranked yet	3 of 5
Chest CT, segmentation		not ranked yet	1 of 4
Spine, bones and joints, segmentation		not ranked yet	3 of 9
Prostate and pelvis, segmentation		not ranked yet	1 of 6
Vessels, segmentation		not ranked yet	1 of 2

Results

Dataset	#	Read by	Score	Failed	Time s
AMOS CT, abdominal organs	4	Dice	0.055	9	19
AVT aortic vessel tree	3	Dice	0.085	9	19
COVID-19 CT Lung and Infection	7	Dice	0.404	4	20
Colorectal-Liver-Metastases, liver and veins	1	Dice	0.949	0	16
CT-ORG	6	Dice	0.089	9	18
CTPelvic1K, CLINIC pelvic bones	5	Dice	0.089	9	22
FLARE22	3	Dice	0.868	0	17
HCC-TACE-Seg	2	Dice	0.841	0	16
HQColon, colon	4	Dice	0.695	0	25
C4KC-KiTS	4	Dice	0.261	12	19
Abdomen CT (Learn2Reg 2020, Task 3)	5	Dice	0.183	0	17
Abdomen CT (Learn2Reg 2021, AbdomenMRCT)	2	Dice	0.902	0	29
LCTSC, thoracic organs at risk	3	Dice	0.858	0	16
CT follow-up of liver metastases, liver	2	Dice	0.623	0	14
PLC-CECT, delayed-phase liver segmentation	5	Dice	0.061	7	17
MSD Liver	4	Dice	0.185	8	20
MSD Pancreas	3	Dice	0.562	6	17
MSD Spleen	2	Dice	0.285	7	17
Spinal-Multiple-Myeloma-SEG, vertebrae and lytic lesions		Dice	0.000	10
PENGWIN CT	5	Dice	0.367	6	19
Pancreas-CT, 8 organs	4	Dice	0.317	6	21
Pancreatic-CT-CBCT-SEG, bowel on the planning CT	4	Dice	0.060	1	14
Pediatric-CT-SEG, organs of children	6	Dice	0.304	7	18
Cardiac CT of pigs, heart and left ventricle		Dice	0.000	10
Prostate-Anatomical-Edge-Cases	1	Dice	0.810	1	20
RCC-AID	2	Dice	0.678	2	16
RibSeg v2, individual ribs	4	Dice	0.014	29
SAROS body regions	5	Dice	0.369	41	29
Spine-Mets-CT-SEG	3	Dice	0.715	0	24
VerSe, vertebrae	3	Dice	0.113	18	16

Source: huggingface.co/MONAI/wholeBody_ct_segmentation
Reference: Wasserthal et al., Radiology: Artificial Intelligence 2023 (the data); Myronenko, BrainLes 2018 (SegResNet)