Pediatric-CT-SEG, organs of children

Chest, abdomen and pelvis CT of one child with up to 28 structures contoured (Pediatric-CT-SEG).

17 test casesthorax, abdomenimage CT

Read before comparing (1)

Cases. The fixed images of PediatricCT's pairs, each on its pair's side of the split. The moving images are left out.

		Accuracy			Cost
#	Method	Dice	Surface Dice	HD95 mm	Time s	RAM GB	GPU GB	Images
1	lungmaskGPUapp2 of 28 structures	0.922 ±0.034	0.785 ±0.067	12.18 ±4.12	20	4.5	4.7
2	KonfAI segmentation app · TotalSegmentator (total), the authors' packageGPUapp21 of 28 structures	0.633 ±0.070	0.495 ±0.060	11.71 ±2.58		19.9	3.8
2	lungmask · lobesGPUapp2 of 28 structures	0.896 ±0.070	0.765 ±0.089	12.22 ±4.10	118	5.1	4.9
44–6	KonfAI segmentation app · TotalSegmentator (total), KonfAI appGPUapp21 of 28 structures	0.624 ±0.071	0.481 ±0.061	16.35 ±5.94	17	2.8	4.9
54–6	KonfAI segmentation app · IMPACTSeg body, KonfAI appGPUapp1 of 28 structures2 failed	0.547 ±0.122	0.395 ±0.089	28.97 ±10.25	15	2.0	6.2
64–6	KonfAI segmentation app · MRSegmentator, KonfAI appGPUapp20 of 28 structures8 failed	0.315 ±0.171	0.242 ±0.133	39.17 ±25.62	31	5.1	9.1
64–6	MONAI wholeBody CTGPUapp19 of 28 structures7 failed	0.304 ±0.138	0.211 ±0.099	17.34 ±3.22	18	4.2	6.2

The order averages each method's ranks on Dice, Surface Dice, HD95 mm, so the first one is not always the best in every column. ± is the 95% interval of the mean over the 17 cases. Two neighbours are tied when a paired Wilcoxon test on the shared cases does not find the upper one better on Dice (5%): the places a method could hold are then shown under its own. The best value of each column is in colour. Click a row to see its images, a column title to sort by it. Time and memory measured on Workstation.

4 results not ranked: an older run, cases still missing, or a failure on every case

		Accuracy			Cost
#	Method	Dice	Surface Dice	HD95 mm	Time s	RAM GB	GPU GB	Machine	Images
	KonfAI segmentation app · TotalSegmentator (total), KonfAI appGPUapp21 of 28 structuresolder run	0.624 ±0.071	0.481 ±0.061	16.35 ±5.94	9	2.6	8.5	Laptop
	lungmaskGPUapp2 of 28 structuresolder run	0.922 ±0.034	0.785 ±0.067	12.18 ±4.12	48	2.8	2.9	Laptop
	lungmask · lobesGPUapp2 of 28 structuresolder run	0.896 ±0.070	0.765 ±0.089	12.22 ±4.10	171	2.8	5.0	Laptop
	MONAI wholeBody CTGPUapp19 of 28 structuresolder run	0.528 ±0.066	0.373 ±0.057	21.68 ±4.70	14	2.8	9.6	Laptop

Dice: how much the organ contours overlap, 1 is perfect. Surface Dice: the share of the contours' surfaces within a tolerance of each other, 1 is perfect. HD95: distance between the contours' surfaces.

Case by case

Measure

Dice	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17
lungmask	0.114	0.072	0.073	0.071	0.077	0.068	0.091	0.077	0.082	0.073	0.078	0.077	0.477	0.079	0.066	0.071	0.082
KonfAI segmentation app · TotalSegmentator (total), the authors' package	0.509	0.525	0.539	0.528	0.592	0.508	0.562	0.662	0.573	0.540	0.567	0.529	0.135	0.524	0.548	0.504	0.535
lungmask · lobes	0.114	0.072	0.073	0.035	0.077	0.068	0.090	0.077	0.082	0.073	0.078	0.077	0.477	0.079	0.066	0.071	0.082
KonfAI segmentation app · TotalSegmentator (total), KonfAI app	0.455	0.526	0.537	0.522	0.587	0.494	0.561	0.651	0.557	0.537	0.566	0.526	0.135	0.522	0.548	0.498	0.540
KonfAI segmentation app · IMPACTSeg body, KonfAI app	0.028	0.023	0.027	0.031	failed	0.018	0.026	failed	0.024	0.030	0.028	0.028	0.268	0.021	0.027	0.016	0.023
KonfAI segmentation app · MRSegmentator, KonfAI app	0.487	0.499	failed	0.509	failed	failed	failed	0.622	0.524	0.515	failed	failed	0.120	failed	0.535	failed	0.512
MONAI wholeBody CT	0.412	0.416	0.434	failed	failed	0.368	failed	failed	0.359	failed	0.463	0.440	failed	0.307	failed	0.354	0.368

Click a score to look at that result: the image, the truth in green and the method's map in red. Underlined scores have their image; for the others you see the case without the result.

About the data

The first test caseClick to look at every case.

Truth: contours by experts up to 28 structures a child, the dataset's contours: 25 organs drawn by hand, the lungs and the bones from a threshold corrected by hand
Source: www.cancerimagingarchive.net/collection/pediatric-ct-seg
doi.org/10.7937/TCIA.X0H0-1706
doi.org/10.1002/mp.15485
Licence: CC BY 4.0

16 cases for tuning, never tested

34ECBB32_8877A5A3/fixed, 379_385/fixed, AA032583_6A2228AD/fixed, 7C7F4414_51733317/fixed, 4EF8F066_8975C78C/fixed, 14403912_CF79E74B/fixed, B722C905_B9DA30EF/fixed, E69AC901_B0308EC6/fixed, 642CA0DF_691FB79F/fixed, 381_8538895B/fixed, B42CD321_F369B022/fixed, 3A4DE292_4934405E/fixed, 627003CD_5AEBE68F/fixed, 47CA11F8_491B83F3/fixed, 211F45E2_29ECC095/fixed, C7D4AEFA_792705D7/fixed