All datasets
AvailableCTMRI

SynthRAD2023 Pelvis Dataset

SynthRAD2023 contains paired pelvic radiotherapy images for synthetic CT generation. Task 1 uses registered MRI and CT images, while Task 2 uses cone-beam CT and CT images acquired at three Dutch university medical centres.

Licence and access

The dataset is distributed under the Creative Commons Attribution-NonCommercial 4.0 International license (CC BY-NC 4.0). Reuse is permitted with attribution for non-commercial purposes; commercial use requires separate permission.

open accesscommercial use restrictedredistribution allowed
Visit official sourcelast verified 2026-09-10
Coverage
pelvissegmentationin vivo
subjects
not verified
images
not verified
formats
NIfTI, XLSX, PNG
origin
Netherlands · 2023
Citation
Thummerer A, van der Bijl E, Maspero M. SynthRAD2023 Grand Challenge dataset: synthetizing computed tomography for radiotherapy. Zenodo. 2023. DOI: 10.5281/zenodo.7260705.
Notes

The complete challenge description reports 540 pelvic image pairs across Task 1 and Task 2: 360 training, 60 validation and 120 test pairs. The linked Zenodo release states that it contains the training data, corresponding to 360 pelvic training pairs. num_subjects and num_images are null because the tasks may contain different patients and a pair contains multiple image volumes. The supplied mask is a binary patient-outline mask rather than a detailed anatomical segmentation. CBCT is described in this note because the current modality schema does not contain a separate CBCT value.

This record is a pointer. Licence and access terms are recorded as they stood on the verification date above; what binds you is the dataset's own licence text. Confirm at the source before use.

Orthopedic AI Research Updates

Monthly research digest, product updates, and clinical AI insights.

Unsubscribe anytime.

SynthRAD2023 Pelvis Dataset, dataset record