TotalSegmentator.md

TotalSegmentator

Dataset Information

TotalSegmentator is the currently largest publicly available dataset for three-dimensional medical image segmentation, including 1204 CT images covering 104 types of anatomical structures. Among them, 1082 images are for training, 57 for validation, and 65 for testing. Unlike most datasets focusing only on partial organs and with limited data volume, TotalSegmentator provides unprecedented data scale and diversity, better adapting to the variable clinical scenarios. In addition, it annotates some structures that are rare in other datasets, offering a solid foundation for model research and optimization.

Dataset Meta Information

Dimensions	Modality	Task Type	Anatomical Structures	Anatomical Area	Number of Categories	Data Volume	File Format
3D	CT	Segmentation	27 organs; 59 bones; 10 muscles; 8 vessels	Entire body	104	1204	.nii.gz

Resolution Details

Dataset Statistics	spacing (mm)	size
min	(1.5, 1.5, 1.5)	(47, 48, 29)
median	(1.5, 1.5, 1.5)	(241, 231, 231)
max	(1.5, 1.5, 1.5)	(499, 430, 851)

Total Number of Slices in the Dataset: 312400 (Statistics for 1203 images, excluding one damaged data s0864)

Label Information Statistics

Label	Anatomical Structure	Occurrence Count	Presence Rate	Mean Volume (cm³)	Volume Variability (cm³)
1	Spleen	870	72.32%	2542.37	176.34
2	Kidney, Right	767	63.76%	432.66	127.42
3	Kidney, Left	786	65.34%	527	126.78
4	Gallbladder	640	53.20%	146.8	17.87
5	Liver	921	76.56%	4162.89	1356.34
6	Stomach	889	73.90%	3175.5	228.41
7	Aorta	1074	89.28%	1447.04	172.47
8	Inferior Vena Cava	961	79.88%	150.43	52.15
9	Portal Vein and Splenic Vein	814	67.66%	65.64	20.74
10	Pancreas	795	66.08%	145.81	59.18
11	Adrenal Gland, Right	811	67.41%	9.88	3.96
12	Adrenal Gland, Left	784	65.17%	13.27	4
13	Lung, Upper Lobe, Left	1008	83.79%	2914.96	794.34
14	Lung, Lower Lobe, Left	900	82.92%	2135.9	640.52
15	Lung, Upper Lobe, Right	885	73.57%	2304.48	731.92
16	Lung, Middle Lobe, Right	865	71.89%	1812.85	330.91
17	Lung, Lower Lobe, Right	975	81.05%	2412.88	744.78
18	Vertebrae, L5	603	50.12%	99.43	61.75
19	Vertebrae, L4	618	51.37%	115.43	62.8
20	Vertebrae, L3	628	52.20%	140.96	62.18
21	Vertebrae, L2	695	57.77%	93.21	55.74
22	Vertebrae, L1	784	65.17%	97.68	50.84
23	Vertebrae, T12	854	70.99%	108.59	44.9
24	Vertebrae, T11	870	72.32%	75	44.16
25	Vertebrae, T10	860	71.49%	85.52	38
26	Vertebrae, T9	833	69.24%	67.92	33.19
27	Vertebrae, T8	830	68.84%	81.54	30.12
28	Vertebrae, T7	731	60.76%	60.16	28.24
29	Vertebrae, T6	722	60.02%	56.8	25.2
30	Vertebrae, T5	759	60.43%	46.19	24.35
31	Vertebrae, T4	735	61.10%	37.59	22.98
32	Vertebrae, T3	759	61.43%	40.2	21.8
33	Vertebrae, T2	721	59.93%	38.82	23.05
34	Vertebrae, T1	701	58.27%	41.32	21.92
35	Vertebrae, C7	688	57.19%	32.37	13.5
36	Vertebrae, C6	554	46.05%	27.45	6.1
37	Vertebrae, C5	348	28.93%	25.66	11.24
38	Vertebrae, C4	249	20.70%	25.07	12.36
39	Vertebrae, C3	228	18.95%	23.21	12.76
40	Vertebrae, C2	243	20.20%	28.83	17.34
41	Esophagus	240	84.37%	17.67	27.35
42	Trachea	788	65.50%	79.61	32.48
43	Heart Myocardium	849	70.57%	264.4	117.96
44	Heart Atrium, Left	834	69.33%	231.52	71.32
45	Heart Ventricle, Left	847	70.41%	245.14	102.92
46	Heart Atrium, Right	849	70.57%	319.19	81.95
47	Heart Ventricle, Right	859	71.40%	388.04	137.71
48	Pulmonary Artery	721	59.93%	155.23	66.18
49	Brain	239	19.87%	1557.75	431.31
50	Iliac Artery, Left	634	52.70%	106.37	15.67
51	Iliac Artery, Right	627	52.12%	129.88	16.1
52	Iliac Vein, Left	642	53.37%	61.56	29.08
53	Iliac Vein, Right	614	51.04%	54.32	24.36
54	Small Bowel	793	65.92%	2584.29	690.8
55	Duodenum	751	62.43%	297.75	48.01
56	Colon	895	74.40%	2046.86	533.74
57	Rib, Left 1st	738	61.35%	211.19	9.25
58	Rib, Left 2nd	749	62.26%	24.73	10.81
59	Rib, Left 3rd	769	63.92%	27.64	11.88
60	Rib, Left 4th	796	66.17%	33.91	14.68
61	Rib, Left 5th	853	70.91%	38.91	15.45
62	Rib, Left 6th	891	74.06%	42.54	17.19
63	Rib, Left 7th	873	72.57%	44.01	18.43
64	Rib, Left 8th	871	72.40%	41.62	15.88
65	Rib, Left 9th	873	72.57%	33.72	15.28
66	Rib, Left 10th	863	71.74%	34.57	12.46
67	Rib, Left 11th	839	69.74%	25.96	7.98
68	Rib, Left 11	869	69.74%	25.96	7.98
69	Rib, Left 12	772	64.17%	10.47	3.14
70	Rib, Right 1	735	61.10%	26.65	9.55
71	Rib, Right 2	746	62.01%	25.05	10.86
72	Rib, Right 3	762	63.34%	28.6	12.04
73	Rib, Right 4	778	64.67%	32.8	15.13
74	Rib, Right 5	825	68.58%	38.53	16.49
75	Rib, Right 6	852	70.82%	43.19	18.29
76	Rib, Right 7	834	69.33%	46.05	19.48
77	Rib, Right 8	843	70.07%	45.69	17.14
78	Rib, Right 9	848	70.70%	39.37	16.07
79	Rib, Right 10	843	70.07%	33.73	12.03
80	Rib, Right 11	847	70.41%	23.15	7.78
81	Rib, Right 12	761	63.26%	12.83	3.08
82	Humerus, Left	642	53.37%	214.21	38.7
83	Humerus, Right	711	59.10%	203.61	38.77
84	Scapula, Left	800	66.50%	327.81	85.73
85	Scapula, Right	779	64.75%	160.98	87.98
86	Clavicle, Left	724	60.18%	62.2	21.76
87	Clavicle, Right	719	59.77%	53.77	23.52
88	Femur, Left	568	47.22%	774.1	158.57
89	Femur, Right	529	43.97%	721.38	161.57
90	Hip, Left	632	52.37%	592.92	359.22
91	Hip, Right	620	51.70%	598.67	359.67
92	Sacrum	603	50.12%	356.18	216.43
93	Face	356	29.59%	1722.87	390.72
94	Gluteus Maximus, Left	582	48.38%	1017.17	443.94
95	Gluteus Maximus, Right	581	48.30%	1084.03	263.22
96	Gluteus Medius, Left	610	50.71%	504.59	246.23
97	Gluteus Medius, Right	599	49.79%	466.93	223.51
98	Gluteus Minimus, Left	565	46.97%	99.59	57.57
99	Gluteus Minimus, Right	556	46.22%	122.36	60.8
100	Autochthon, Left	1115	92.68%	796.14	285.7
101	Autochthon, Right	1113	92.52%	808.46	279.46
102	Iliopsoas, Left	818	68.00%	571.27	200.07
103	Iliopsoas, Right	854	68.50%	635.86	194.32
104	Ureter, Right (Distal End)	829	68.57%	44.76	14.34

Visualization

Visualization using itk-snap.

File Structure

After unzipping the official compressed package, the file structure is as follows: it contains a meta.csv file and multiple sxxxx subdirectories. In meta.csv, each line represents an imaging data, listing the picture ID (image_id), age (age), gender (gender), institutional code (institute), body part or examination range (study_type), and dataset division (split). These data cover CT scans of multiple body parts, sourced from different institutions, and are divided into training, validation, and testing subsets. Each sxxxx subdirectory contains a segmentations folder and a ct.nii.gz file.

TotalSegmentator
│
├── meta.csv
│
├── s0000
│   ├── segmentations
│   │   ├── adrenal_gland_left.nii.gz
│   │   ├── adrenal_gland_right.nii.gz
│   │   ├── aorta.nii.gz
│   │   └── ...
│   ├── ct.nii.gz
│
├── s0001
├── s0002
├── s0003
└── ...

Authors and Institutions

Jakob Wasserthal (Department of Radiology and Nuclear Medicine, University Hospital Basel, Switzerland)

Source Information

Official Website: https://github.com/wasserth/TotalSegmentator

Download Link: https://zenodo.org/record/6802614

Article Address: https://pubs.rsna.org/doi/10.1148/ryai.230024

Data Public Release Date: 2022-07

Citation

@article{totalsegmentator,
    author = {Wasserthal, Jakob and Breit, Hanns-Christian and Meyer, Manfred T. and Pradella, Maurice and Hinck, Daniel and Sauter, Alexander W. and Heye, Tobias and Boll, Daniel T. and Cyriac, Joshy and Yang, Shan and Bach, Michael and Segeroth, Martin},
    title = {TotalSegmentator: Robust Segmentation of 104 Anatomic Structures in CT Images},
    journal = {Radiology: Artificial Intelligence},
    volume = {5},
    number = {5},
    pages = {e230024},
    year = {2023}
}

Original introduction article is here.