Datasets:

Angelou0516
/

BIDMC

license: other
license_name: attribution
tags:
  - medical-imaging
  - segmentation
  - prostate
  - mri
  - multi-site
task_categories:
  - image-segmentation
pretty_name: Multi-Site Prostate MRI Segmentation (BIDMC / FedDG)
dataset_info:
  features:
    - name: case_id
      dtype: string
    - name: site
      dtype: string
    - name: num_slices
      dtype: int32
    - name: image
      dtype: image
    - name: mask
      dtype: image
    - name: overlay
      dtype: image
  splits:
    - name: train
      num_bytes: 23756040
      num_examples: 146
  download_size: 23762490
  dataset_size: 23756040
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Multi-Site Prostate MRI Segmentation Dataset

T2-weighted prostate MRI volumes from 6 clinical sites, curated for domain generalization research in medical image segmentation.

Dataset Details

Property	Value
Modality	MRI (T2-weighted)
Organ	Prostate (whole gland)
Total cases	116
Format	NIfTI (.nii.gz)

Sites

Site	Institution	Cases
RUNMC	Radboud Univ. Nijmegen Medical Centre	30
BMC	Boston Medical Center	30
I2CVB	Init. for Collaborative CV Benchmarking	19
UCL	University College London	13
BIDMC	Beth Israel Deaconess Medical Center	12
HK	Haukeland University Hospital	12

File Structure

Each site folder contains paired volumes:

CaseXX.nii.gz — T2-weighted MRI volume
CaseXX_segmentation.nii.gz — Binary prostate mask

Citation

@inproceedings{liu2021feddg,
  title={FedDG: Federated Domain Generalization on Medical Image Segmentation via Episodic Learning in Continuous Frequency Space},
  author={Liu, Quande and Chen, Cheng and Qin, Jing and Dou, Qi and Heng, Pheng-Ann},
  booktitle={CVPR},
  year={2021}
}