Dataset Viewer
Auto-converted to Parquet Duplicate
patient_id
stringlengths
18
18
scanner
stringclasses
2 values
num_slices
int64
15
38
image
imagewidth (px)
256
432
segmentation
imagewidth (px)
256
432
overlay
imagewidth (px)
256
432
Prostate3T-01-0001
3T-Siemens
15
Prostate3T-01-0002
3T-Siemens
20
Prostate3T-01-0003
3T-Siemens
24
Prostate3T-01-0004
3T-Siemens
20
Prostate3T-01-0005
3T-Siemens
20
Prostate3T-01-0006
3T-Siemens
20
Prostate3T-01-0007
3T-Siemens
20
Prostate3T-01-0008
3T-Siemens
15
Prostate3T-01-0009
3T-Siemens
17
Prostate3T-01-0010
3T-Siemens
20
Prostate3T-01-0011
3T-Siemens
20
Prostate3T-01-0012
3T-Siemens
20
Prostate3T-01-0013
3T-Siemens
15
Prostate3T-01-0014
3T-Siemens
20
Prostate3T-01-0015
3T-Siemens
20
Prostate3T-01-0016
3T-Siemens
20
Prostate3T-01-0017
3T-Siemens
20
Prostate3T-01-0018
3T-Siemens
20
Prostate3T-01-0019
3T-Siemens
18
Prostate3T-01-0020
3T-Siemens
20
Prostate3T-01-0021
3T-Siemens
20
Prostate3T-01-0022
3T-Siemens
20
Prostate3T-01-0023
3T-Siemens
20
Prostate3T-01-0024
3T-Siemens
20
Prostate3T-01-0025
3T-Siemens
19
Prostate3T-01-0026
3T-Siemens
20
Prostate3T-01-0027
3T-Siemens
20
Prostate3T-01-0028
3T-Siemens
15
Prostate3T-01-0029
3T-Siemens
20
Prostate3T-01-0030
3T-Siemens
20
ProstateDx-01-0001
1.5T-Philips
32
ProstateDx-01-0005
1.5T-Philips
30
ProstateDx-01-0006
1.5T-Philips
32
ProstateDx-01-0011
1.5T-Philips
32
ProstateDx-01-0013
1.5T-Philips
32
ProstateDx-01-0014
1.5T-Philips
36
ProstateDx-01-0019
1.5T-Philips
32
ProstateDx-01-0021
1.5T-Philips
32
ProstateDx-01-0023
1.5T-Philips
32
ProstateDx-01-0028
1.5T-Philips
32
ProstateDx-01-0035
1.5T-Philips
38
ProstateDx-01-0038
1.5T-Philips
32
ProstateDx-01-0042
1.5T-Philips
32
ProstateDx-01-0043
1.5T-Philips
32
ProstateDx-01-0044
1.5T-Philips
32
ProstateDx-01-0052
1.5T-Philips
32
ProstateDx-01-0054
1.5T-Philips
32
ProstateDx-01-0055
1.5T-Philips
34
ProstateDx-01-0056
1.5T-Philips
32
ProstateDx-01-0058
1.5T-Philips
32
ProstateDx-01-0059
1.5T-Philips
32
ProstateDx-01-0060
1.5T-Philips
32
ProstateDx-01-0063
1.5T-Philips
32
ProstateDx-01-0071
1.5T-Philips
32
ProstateDx-01-0073
1.5T-Philips
35
ProstateDx-01-0074
1.5T-Philips
34
ProstateDx-01-0075
1.5T-Philips
34
ProstateDx-01-0076
1.5T-Philips
32
ProstateDx-01-0080
1.5T-Philips
32
ProstateDx-01-0082
1.5T-Philips
32

NCI-ISBI 2013 — Automated Segmentation of Prostate Structures

T2-weighted MRI dataset for prostate zonal segmentation (Peripheral Zone and Central Gland).

Dataset Details

Field Value
Modality T2-weighted MRI (1.5T + 3T)
Body part Prostate
Classes 2 — Peripheral Zone (PZ), Central Gland (CG)
Subjects 80 (60 train / 10 leaderboard / 10 test)
Format DICOM images + NRRD segmentations
License CC BY 3.0

Structure

images/
  Training/       # 60 cases (30 Prostate3T + 30 ProstateDx)
  Leaderboard/    # 10 cases
  Test/           # 10 cases
segmentations/
  Training/       # 60 NRRD masks
  Leaderboard/    # 10 NRRD masks
  Test/           # 10 NRRD masks

Each patient directory under images/ contains DICOM slices (~15-38 per case). Segmentation masks are NRRD files with matching patient IDs.

  • Prostate3T-*: acquired at 3T (Siemens TIM, surface coil)
  • ProstateDx-*: acquired at 1.5T (Philips Achieva, endorectal coil)

Segmentation Labels

Label Structure
0 Background
1 Central Gland (CG)
2 Peripheral Zone (PZ)

Source

Citation

@misc{bloch2015nciisbi,
  author = {Bloch, N. and Madabhushi, A. and Huisman, H. and Freymann, J. and Kirby, J. and Grauer, M. and Enquobahrie, A. and Jaffe, C. and Clarke, L. and Farahani, K.},
  title = {NCI-ISBI 2013 Challenge: Automated Segmentation of Prostate Structures},
  year = {2015},
  publisher = {The Cancer Imaging Archive},
  doi = {10.7937/K9/TCIA.2015.zF0vlOPv}
}
Downloads last month
1,330