Skip to content

Supported Datasets

ITKIT provides dataset restructuring scripts to convert datasets from their official release formats to a consistent ITKIT structure. This enables unified API usage across different datasets.

Conversion Scripts

For each supported dataset, you can find conversion scripts at: itkit/dataset/<dataset_name>/convert_<format>.py

Dataset List

1. AbdomenCT-1K

Description: Large-scale abdominal organ segmentation dataset.

Organs: Multiple abdominal organs

Modality: CT

Reference: J. Ma et al., "AbdomenCT-1K: Is Abdominal Organ Segmentation a Solved Problem?," in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 10, pp. 6695-6714, 1 Oct. 2022.

DOI: 10.1109/TPAMI.2021.3100536

Link: IEEE Xplore

2. BraTS 2024

Description: Brain Tumor Segmentation Challenge - Glioma segmentation on post-treatment MRI.

Organs: Brain, tumor subregions

Modality: MRI (multi-sequence)

Reference: Maria Correia de Verdier, et al., "The 2024 Brain Tumor Segmentation (BraTS) Challenge: Glioma Segmentation on Post-treatment MRI," arXiv preprint arXiv:2405.18368, 2024.

arXiv: 2405.18368

3. CT-ORG

Description: Multiple organ segmentation in computed tomography.

Organs: Liver, bladder, lungs, kidneys, bone, brain

Modality: CT

Reference: Rister, B., Yi, D., Shivakumar, K. et al. CT-ORG, a new dataset for multiple organ segmentation in computed tomography. Sci Data 7, 381 (2020).

DOI: 10.1038/s41597-020-00715-8

Link: Nature

4. CTSpine1K

Description: Large-scale dataset for spinal vertebrae segmentation.

Organs: Vertebrae

Modality: CT

Reference: Yang Deng, Ce Wang, Yuan Hui, et al. CtSpine1k: A large-scale dataset for spinal vertebrae segmentation in computed tomography. arXiv preprint arXiv:2105.14711 (2021).

arXiv: 2105.14711

5. FLARE 2022

Description: Fast and Low-resource semi-supervised Abdominal oRgan sEgmentation.

Organs: Liver, kidneys, spleen, pancreas

Modality: CT

Reference: Jun Ma, et al., Unleashing the Strengths of Unlabeled Data in Pan-cancer Abdominal Organ Quantification: the FLARE22 Challenge. arXiv preprint arXiv:2308.05862, 2023.

arXiv: 2308.05862

6. FLARE 2023

Description: Fast, Low-resource, and Accurate Organ and Pan-cancer Segmentation in Abdomen CT.

Organs: Multiple abdominal organs and tumors

Modality: CT

Reference: Jun Ma, Bo Wang (Eds.). Fast, Low-resource, and Accurate Organ and Pan-cancer Segmentation in Abdomen CT: MICCAI Challenge, FLARE 2023, Held in Conjunction with MICCAI 2023, Vancouver, BC, Canada, October 8, 2023, Proceedings. Lecture Notes in Computer Science. Springer, Cham, 2024.

DOI: 10.1007/978-3-031-58776-4

Link: Springer

7. ImageTBAD

Description: 3D Computed Tomography Angiography Image Dataset for Automatic Segmentation of Type-B Aortic Dissection.

Organs: Aorta

Modality: CTA

Reference: Yao Z, Xie W, Zhang J, Dong Y, Qiu H, Yuan H, Jia Q, Wang T, Shi Y, Zhuang J, Que L, Xu X and Huang M (2021) ImageTBAD: A 3D Computed Tomography Angiography Image Dataset for Automatic Segmentation of Type-B Aortic Dissection. Front. Physiol. 12:732711.

DOI: 10.3389/fphys.2021.732711

Link: Frontiers

8. KiTS23

Description: Kidney and Kidney Tumor Segmentation Challenge.

Organs: Kidneys, renal tumors, renal cysts

Modality: CT (corticomedullary phase)

References:

  1. Nicholas Heller, Fabian Isensee, Dasha Trofimova, et al. The KiTS21 Challenge: Automatic segmentation of kidneys, renal tumors, and renal cysts in corticomedullary-phase CT. arXiv:2307.01984 [cs.CV], 2023.
  2. Nicholas Heller, Fabian Isensee, Klaus H. Maier-Hein, et al. The state of the art in kidney and kidney tumor segmentation in contrast-enhanced CT imaging: Results of the KiTS19 challenge. Medical Image Analysis, Vol. 67, Article 101821, 2021.

Website: kits-challenge.org

DOI: 10.1016/j.media.2020.101821

9. LUNA16

Description: Lung Nodule Analysis - automatic detection of pulmonary nodules.

Organs: Lungs, nodules

Modality: CT

Reference: Arnaud Arindra Adiyoso Setio, Alberto Traverso, Thomas de Bel, et al. Validation, comparison, and combination of algorithms for automatic detection of pulmonary nodules in computed tomography images: The LUNA16 challenge. Medical Image Analysis, Vol. 42, pp. 1–13, 2017.

DOI: 10.1016/j.media.2017.06.015

Link: ScienceDirect

10. TCGA

Description: The Cancer Genome Atlas - comprehensive cancer genomics dataset.

Organs: Various cancer types

Modality: Multiple

Website: cancer.gov/TCGA

11. TotalSegmentator

Description: Robust segmentation of 104 anatomic structures in CT images.

Organs: 104 anatomical structures

Modality: CT

Reference: Wasserthal Jakob, et al. TotalSegmentator: Robust Segmentation of 104 Anatomic Structures in CT Images. Radiology: Artificial Intelligence, 5, 5, 2023.

DOI: 10.1148/ryai.230024

Link: RSNA

12. LiTS

Description: Liver Tumor Segmentation Benchmark.

Organs: Liver, liver tumors

Modality: CT

Reference: Bilic, Patrick and Christ, Patrick and Li, Hongwei Bran and Vorontsov, Eugene and Ben-Cohen, Avi and Kaissis, Georgios and Szeskin, Adi and Jacobs, Colin and Mamani, Gabriel Efrain Humpire and Chartrand, Gabriel and others. The liver tumor segmentation benchmark (lits). Medical Image Analysis, volume 84, 2023, 102680.

DOI: 10.1016/j.media.2022.102680

Link: ScienceDirect

13. MSD

Description: Medical Segmentation Decathlon

Organs: Various (10 tasks)

Modality: Various (CT, MRI)

Reference: Antonelli, M., Reinke, A., Bakas, S. et al. The Medical Segmentation Decathlon. Nat Commun 13, 4128 (2022).

DOI: 10.1038/s41467-022-30695-9

Link: Nature Communications

Example: Converting AbdomenCT-1K

# Navigate to dataset conversion directory
cd itkit/dataset/AbdomenCT-1K/

# Run conversion script
python convert_official.py \
    --input /path/to/official/dataset \
    --output /path/to/itkit/format

General Workflow

  1. Download official dataset from the source
  2. Locate conversion script in itkit/dataset/<dataset_name>/
  3. Run conversion script with appropriate paths
  4. Verify structure matches ITKIT format (image/ and label/ folders)
  5. Use with ITKIT tools for preprocessing and training

Custom Dataset Preparation

If your dataset is not in the list, you can manually convert it to ITKIT format:

  1. Create image/ and label/ folders
  2. Place image files in image/
  3. Place corresponding label files in label/ with matching names
  4. Use any supported format (.mha, .nii.gz, etc.)
  5. Optionally create meta.json with dataset information

Contributing New Datasets

If you've created a conversion script for a new dataset, please consider contributing it to ITKIT:

  1. Create conversion script following the existing pattern
  2. Add dataset documentation
  3. Submit a pull request

See Contributing Guide for details.