The Dataset Directory
The first table is the selected, source-reviewed directory generated from data/datasets.csv. Counts describe the named release and unit; a frame, image, examination, slide, and patient are not interchangeable. Source-review scope and modality keys are documented in data/README.md. Entries reviewed for the remaining chapters on 22 September 2026 complement the earlier chest-radiography entries.
| Modality | Dataset | Size / release | Supervision | Access / terms | Caveats | Source |
|---|---|---|---|---|---|---|
| cxr | ChestX-ray14 (NIH) | 112,120 images, 30,805 patients | 14 findings (NLP-mined) | Public research release; consult NIH terms | Report-mined labels; single US center; release documentation governs splits | Source |
| cxr | CheXpert | 224,316 images; 65,240 patients | 14 findings, uncertainty labels | Access request and Stanford dataset agreement | Single US center; report-mined training labels; expert evaluation | Source |
| cxr | MIMIC-CXR | 377,110 images; 227,835 studies (v2 release) | DICOM images and free-text reports | PhysioNet credentialing; training; data-use agreement | Single Boston center; JPG derivative overlaps this corpus | Source |
| cxr | PadChest | 160,868 images | 174 findings; 27% manually labeled | Research-use agreement; access request | Spanish hospital; manual and automated annotations; restricted redistribution | Source |
| cxr | VinDr-CXR | 18,000 images | 22 local findings with boxes; 6 global categories | PhysioNet credentialing and data-use agreement | Adult PA images from two Vietnamese hospitals; 15,000 train / 3,000 test | Source |
| ct | LIDC-IDRI | 1,018 CT scans | Multi-reader lung nodule annotations | TCIA collection terms | Reader disagreement is part of the reference | Source |
| ct | TotalSegmentator CT v2 | 1,228 CT scans | 117 anatomical structures | CC BY 4.0; release terms | Pin label map and release; CT dataset differs from MR release | Source |
| mri | fastMRI | Release-specific knee and brain collections | Raw k-space and reconstruction references | Registration and dataset agreement | Do not combine release counts; preserve coil and acquisition metadata | Source |
| mri | BraTS | Annual task-specific releases | Defined brain-tumor subregions | Challenge-specific access and terms | Pin year/task; preprocessing and populations vary | Source |
| us | EchoNet-Dynamic | 10,030 apical four-chamber videos | EF and selected ventricular tracings | Research data agreement | Selected clips; not a complete live-acquisition dataset | Source |
| us | CAMUS | 500 patients | Two- and four-chamber contours at defined phases | Challenge access and terms | Keep all views and phases from a patient together | Source |
| mammo | VinDr-Mammo v1.0.0 | 5,000 four-view examinations; 20,000 images | Breast-level assessments and finding annotations | PhysioNet credentialing and data-use agreement | Full-field digital mammography; not DBT | Source |
| mammo | CBIS-DDSM | Curated mass and calcification subsets | Lesion annotations and pathology-linked labels | TCIA collection terms | Digitized film; do not describe as modern digital acquisition | Source |
| nm | FDG-PET-CT-Lesions / autoPET source | Release-specific whole-body PET/CT collection | FDG-avid lesion segmentations | TCIA collection and challenge terms | Pin release and split; FDG evidence does not transfer to other tracers | Source |
| ophtho | EyePACS diabetic-retinopathy competition | Competition-specific train and test releases | DR severity grades | Kaggle competition and data terms | Keep both eyes and repeated captures grouped; inspect quality exclusions | Source |
| path | CAMELYON16 / CAMELYON17 | Separate challenge releases | Lymph-node metastasis; release-specific region and case labels | Challenge release terms | Do not pool without checking patient grouping and overlap | Source |
| video | Cholec80 | 80 cholecystectomy videos | Surgical phase and instrument annotations | Research access agreement | Split by procedure; distinguish causal from offline evaluation | Source |
| video | Kvasir-SEG | 1,000 polyp images | Segmentation masks | Dataset-specific terms; inspect source release | Selected still images do not validate live event detection | Source |
| derm | ISIC archive | Collection-specific releases | Lesion metadata and task-specific labels | Image- and collection-specific licenses | Keep repeated lesion images together; verify patient grouping | Source |
| ecg | PTB-XL v1.0.3 | Versioned 12-lead ECG waveform collection | Clinical ECG statements | PhysioNet release license and attribution | Native waveforms; rendered derivatives are not naturally scanned ECGs | Source |
Community dataset sources
The larger directory below is a snapshot of the community medical-image dataset spreadsheet, captured on 22 September 2026. It combines the organizer and participant dataset tabs, removes duplicate dataset–modality listings and submissions that are not identifiable datasets, and retains worksheet-row provenance. The sheet also has a model tab, which is not a dataset source. A dataset can appear in more than one modality group when the sheet lists different image types or releases.
These links and modality/target descriptions are as supplied by the spreadsheet, not an independent review of release contents, current availability, licensing, or suitability. “Paper” means the sheet supplied a publication link but no dataset link. Consult the original source before downloading or using any dataset; the selected directory above remains the more carefully reviewed starting point.
CT (68)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| AbdomenAtlas | Abdomen | CT | Abdominal organs | Dataset (organizers row 30) |
| AbdomenCT-1K | Abdomen | CT | Liver, Kidneys, Pancreas, Spleen | Dataset (organizers row 2) |
| Adrenal-ACC-Ki67-Seg | Abdomen | CT | Adrenocortical carcinoma | Dataset (organizers row 3) |
| AeroPath | Chest | CT | airways | Dataset (organizers row 155) |
| AMOS | Abdomen | CT | Abdominal organs | Dataset (organizers row 4) |
| AortaSeg24 | Abdomen | CT | Aorta | Dataset (participants row 54) |
| Aortic-Dissection | Chest | CTA | true lumina, false lumina | Dataset (organizers row 168) |
| ATM22 | Chest | CT | Airway trees | Dataset (organizers row 51) |
| BTCV | Abdomen | CT | abdominal organs | Dataset (participants row 3) |
| CHAOS | Abdomen | CT | Abdomen organs | Dataset (organizers row 62) |
| Colorectal-Liver_Metastases | Abdomen | CT | Liver, Colorectal Liver Metastases | Dataset (organizers row 66) |
| COVID-19-20 | Chest | CT | COVID-19 lesions | Dataset (organizers row 7) |
| COVID-19-20-CTSEG | Chest | CT | COVID-19 Infections, left lung, right lung | Dataset (organizers row 8) |
| COVID-19-CT-Seg-Benchmark | Chest | CT | Lungs | Dataset (organizers row 19) |
| CT Lymph Nodes | Chest, Abdomen | CT | lymph nodes | Dataset (organizers row 20) |
| CT_ORG | Abdomen | CT | Abdominal organs | Dataset (participants row 4) |
| CURVAS | Abdomen | CT | pancrease, kidney, liver | Dataset (organizers row 165) |
| DeepLesion-MedSAM2 | Whole-body lesions | CT | Lesion | Dataset (organizers row 173) |
| FDG-PET-CT-Lesions (autoPET) | whole body | CT | tumor | Dataset (organizers row 27) |
| GammaKnife-Hippocampal | Brain | CT | brain organ | Dataset (organizers row 76) |
| HCC-TACE-Seg | Abdomen | CT | Tumor | Dataset (organizers row 12) |
| INSTANCE22 | Head | CT | Intracranial Hemorrhage | Dataset (organizers row 10) |
| Kipa22 | Abdomen | CT | Kidney, Vessel, Tumor | Dataset (participants row 64) |
| KiTS23 | Abdomen | CT | Kidney Lesion | Dataset (organizers row 11) |
| LCTSC | Chest | CT | Esophagus, Heart, Lungs, Spinal cord | Dataset (organizers row 24) |
| LIDC-IDRI | Chest | CT | Lung cancer | Dataset (organizers row 88) |
| LiTS | Abdomen | CT | Liver tumor | Dataset (participants row 5) |
| LNDb | Lung | CT | Lung cancer | Dataset (participants row 7) |
| LongCIU | Chest | CT | ground glass opacities (GGO), consolidations | Dataset (organizers row 135) |
| LUNA 16 | Chest | CT | Lung nodule | Dataset (organizers row 89) |
| Mediastinal-Lymph-Node-SEG | Chest | CT | lymph nodes | Dataset (organizers row 21) |
| MedSeg_IVC | — | CT | — | Dataset (organizers row 171) |
| MedSeg_Liver-Segments | — | CT | — | Dataset (organizers row 170) |
| MSD-ColonCancer | Abdomen | CT | Colon Cancer Primaries | Dataset (organizers row 6) |
| MSD-HepaticVessels | Vessel | CT | Hepatic vessels and tumour | Dataset (organizers row 14) |
| MSD-LiverTumor | Abdomen | CT | Liver and tumour | Dataset (organizers row 13) |
| MSD-LungTumor | Lung | CT | Lung tumor | Dataset (organizers row 15) |
| MSD-Pancreastumor | Abdomen | CT | pancreas and tumors | Dataset (organizers row 22) |
| MSD-Spleen | Abdomen | CT | Spleen | Dataset (organizers row 95) |
| MSWAL | Abdomen | CT | gallstones, kidney stones, liver tumors, kidney tumors, pancreatic cancer, liver cysts, and kidney cysts. | Dataset (participants row 74) |
| NIH-Pancreas | Pancreas | CT | Pancreas | Dataset (participants row 48) |
| NSCLC Radiogenomics | Chest | CT | tumor | Dataset (organizers row 16) |
| NSCLC-PleuralEffusion | Chest | CT | PleuralEffusion | Dataset (organizers row 17) |
| NSCLC-Radiomics | Chest | CT | tumor | Dataset (organizers row 18) |
| OPC-Radiomics | Head and Neck | CT | Oropharynx tumor | Dataset (organizers row 98) |
| Pancreatic-CT-CBCT-SEG | Abdomen | CT | Organ at risk | Dataset (organizers row 100) |
| PANORAMA | Abdomen | CT | Pancreas | Dataset (organizers row 128) |
| Parse 2022 | Chest | CT | Pulmonary Artery | Dataset (organizers row 103) |
| Pediatric-CT-SEG | Chest, Abdomen, Pelvis | CT | Abdomen, Chest, Pelvic organs | Dataset (organizers row 104) |
| PENGWIN | Pelvis | CT | Pelvic fracture | Dataset (organizers row 160) |
| Prostate Anatomical Edge Cases | Pelvis | CT | Prostate, rectum, bladder, and bilateral femoral heads | Dataset (organizers row 105) |
| quan | Lung | CT | Lung | Dataset (participants row 6) |
| RibSeg | Abdomen | CT | Ribs | Dataset (participants row 50) |
| Seg.A. | Vessel | CTA | Aorta | Dataset (organizers row 5) |
| SegRap | Head | CT | Head-neck cancer | Dataset (organizers row 23) |
| SegTHOR | Heart | CT | Thoracic organs | Dataset (participants row 8) |
| SpineMets | Abdomen | CT | Spine Mets | Dataset (participants row 60) |
| StructSeg | Head and Neck | CT | 22 OAR Head & Neck Nasopharynx Cancer 6 OAR Lung | Dataset (participants row 46) |
| STS-3D-Tooth | Oral | CBCT | tooth | Dataset (organizers row 143) |
| TCGA-HNSC | Head and Neck | CT | Head-neck cancer | Dataset (organizers row 108) |
| ToothFairy 2 | Head and Neck | CT | Dental Structures | Dataset (participants row 66) |
| TopCoW24 | Brain | CTA | vessel | Dataset (organizers row 124) |
| TotalSegmentator | WholeBody | CT | Whole body organs | Dataset (organizers row 25) |
| USL | Chest, Abdomen, Pelvis | CT | lesion | Dataset (organizers row 113) |
| VERSE20 | Spine | CT | Vertebrae | Dataset (organizers row 115) |
| VESSEL12 | Chest | CT | Lung Vessel | Dataset (organizers row 116) |
| WAW-TCE | Abdomen | CT | Liver Tumor | Dataset (organizers row 158) |
| WORD | Abdomen | CT | Abdomen organs | Dataset (organizers row 117) |
MRI (60)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| ACDC | Chest | MR | Right Ventricle, Left Ventricle, Myocardium | Dataset (organizers row 31) |
| AMOS | Abdomen | MR | Abdominal organs | Dataset (organizers row 28) |
| Aphasia Recovery Cohort (ARC) Dataset | Head and Neck | MR | brain lesion | Dataset (participants row 23) |
| Atlas Bourgogne | Abdomen | MR | Liver, Tumor | Dataset (participants row 61) |
| BONBID-HIE | Head | MR | Hypoxic ischemic encephalopathy lesions | Dataset (organizers row 147) |
| Brain TR-GammakKnife | Brain | MR | brain lesion | Dataset (organizers row 56) |
| BrainMetShare | Head and Neck | MR | Brain tumors | Dataset (participants row 24) |
| BrainPTM | Brain | MR-T1 | White matter tracts | Dataset (organizers row 57) |
| BraTS | Head | MR-T1, T1Gd, T2, FLAIR | Brain Lesion | Dataset (organizers row 29) |
| CAP | Heart | MR | Left Ventricle | Dataset (participants row 58) |
| cDEMRIS | Heart | CMR | Left atrial wall | Dataset (participants row 2) |
| CHAOS | Abdomen | MR | Abdomen organs | Dataset (organizers row 32) |
| CrossMoDA | Brain | MR-T1-T2 | Tumor, Cochlea | Dataset (organizers row 33) |
| Duke-Liver | Abdomen | MR | Liver | Dataset (organizers row 72) |
| EMIDEC | Heart | MRI | Left-ventricular-cavity normal-myocardium myocardial infarction no-reflow | Dataset (organizers row 166) |
| FaBiAN v2.0 | — | MR | — | Dataset (organizers row 164) |
| FeTA | Brain | MR | Brain organs | Dataset (organizers row 73) |
| Figshare Brain Tumor Dataset | Brain | MR-T1ce | meningioma; glioma; pituitary tumor | Dataset (organizers row 75) |
| GammaKnife-Hippocampal | Brain | MR | brain organ | Dataset (organizers row 77) |
| GBM-Reservoir | Brain | MR (FLAIR, T1, T1ce, T2) | brain tumor (Enhancing tumor, Tumor core, Complete tumor extent/whole tumor) | Dataset (organizers row 167) |
| Hippo | hippocampus | MR | hippocampal formation subfields (ERC, SUB, CA1, DG-CA4, CA2, CA3, the hippocampal tail) | Dataset (organizers row 139) |
| HNTSMRG | Head | MR | Tumor | Dataset (organizers row 34) |
| HVSMR | Heart | MR | blood pool, ventricular myocardium | Dataset (organizers row 130) |
| I2CVB | Pelvis | MR | Prostate Gland, Peripheral Zone, Central Gland, CaP | Dataset (organizers row 83) |
| IBD | Abdomen | MRE | stomach, duodenum, small intestine, appendix, cecum, ascending colon, transverse colon, descending colon, sigmoid colon, rectum | Dataset (organizers row 134) |
| ISBILesion2015 | Head and Neck | MR | MS Lesions | Dataset (participants row 45) |
| ISLES | Brain | MR-DWI-ADC, FLAIR | Ischemic stroke lesion | Dataset (organizers row 35) |
| LGGFLair | Head and Neck | FLAIR | low-grade glioma brain tumor | Dataset (participants row 13) |
| LGGMRISeg | Head and Neck | MR | Tumor | Dataset (participants row 57) |
| LLD-MMRI-MedSAM2 | Abdomen | MRI | Liver lesion | Dataset (organizers row 174) |
| MAMA-MIA | breast | MR | Breast Cancer lesions (primary tumors and non-mass-enhanced regions) | Dataset (organizers row 133) |
| MedSeg_Lateral-Ventricles | — | MR | — | Dataset (organizers row 169) |
| Meningioma-SEG-CLASS | Brain | MR | Tumor (Meningioma) | Dataset (organizers row 91) |
| MMs-20-21 | Chest | MR | LV, Myocardium, RV | Dataset (organizers row 93) |
| MSD-Hippocampus | Head | MR | Hippocampus | Dataset (organizers row 96) |
| MSD-LeftAtrium | Chest | MR | Left Atrium | Dataset (organizers row 36) |
| MSD-Prostate | Pelvis | MR (T2, ADC) | Prostate | Dataset (organizers row 37) |
| NCI-ISBI | Pelvis | MR-T2 | Prostate | Dataset (organizers row 38) |
| OASIS | Brain | MR | cerebral spinal fluid, gray matter, white matter | Dataset (organizers row 131) |
| PanSegData | Abdomen | MR | Pancreas | Dataset (organizers row 39) |
| PI-CAI | Pelvis | MR | Prostate Cancer | Dataset (organizers row 109) |
| Pituitary_tumor | Head and neck | MR | Pituitary Neuroendocrine Tumor, carotid | Dataset (organizers row 148) |
| PROMISE12 | Pelvis | MR-T2 | Prostate | Dataset (organizers row 40) |
| PROSTATE-MRI | Prostate | MRI Histopathology | Prostate tumor | Dataset (participants row 69) |
| ProstateX | Prostate | MR | Prostate Lesion | Dataset (participants row 49) |
| QIN-PROSTATE | Pelvis | MR | Prostate and lesion | Dataset (organizers row 41) |
| ReMIND | Head and Neck | MR | Brain Cancer | Dataset (participants row 67) |
| RESECT | Brain | MR | tumor | Dataset (organizers row 156) |
| SCD | Heart | MR | Heart anatomy | Dataset (participants row 25) |
| SegThy | Neck | MRI | Thyroid, Carotid, Jugular Vein | Dataset (organizers row 127) |
| SPIDER | Bone | MR | IVD | Dataset (organizers row 42) |
| SpineMR | Bone | MR | Spine | Dataset (organizers row 45) |
| Stanford Knee MRI Multi-Task Evaluation (SKM-TEA) Dataset | Knee | MR | Cartilage (Femur-Tibia-Patella) & Meniscus | Dataset (participants row 42) |
| TCIA_CervicalCancer | Pelvis | MR | Tumor | Dataset (organizers row 46) |
| TOM500 | Head | MR | optic nerve, orbital fat, lacrimal gland, eyeball, superior rectus and levator palpebrae superioris complex, inferior rectus, medial rectus, lateral rectus, superior oblique | Dataset (organizers row 146) |
| TopCoW24 | Brain | MRA | vessel | Dataset (organizers row 125) |
| TotalSegmentatorMR | WholeBody | MR | Whole body organs | Dataset (organizers row 47) |
| TrackRAD | Whole-body | MRI | Lesion | Dataset (organizers row 129) |
| UCSF-PDGM | Brain | MR | brain lesion | Dataset (organizers row 110) |
| WMH | Brain | MR-T1, FLAIR | White matter hyper-intensities | Dataset (organizers row 48) |
Ultrasound (20)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| AbdominalUS | Abdomen | Ultrasound | Organs | Dataset (participants row 32) |
| ACOUSLIC | Abdomen | US | Abdominal circumference | Dataset (organizers row 159) |
| BUID | Breast | Ultrasound | Tumors | Dataset (participants row 35) |
| BUS-UCLM | Breast | Ultrasound | breast lesions | Dataset (organizers row 140) |
| BUSI | Chest | US | benign breast cancer malignant breast cancer | Dataset (organizers row 59) |
| CAMUS | Heart | US | left ventricle, myocardium, left atrium | Dataset (organizers row 50) |
| CT2US | Abdomen | Ultrasound | Liver | Dataset (participants row 30) |
| CT2USforKidneySeg | Abdomen | US | kidney | Dataset (organizers row 69) |
| DDTI | Head and Neck | Ultrasound | Thyroid | Dataset (participants row 33) |
| hc18 | Head | US | fetal head circumference | Dataset (organizers row 79) |
| HMC-QU | Chest | US | left ventricle | Dataset (organizers row 80) |
| Leg-3D-US | Low-limb leg | US | Soleus, Gastrocnemius Medialis, Gastrocnemius Lateralis | Dataset (organizers row 49) |
| Micro-Ultrasound-Prostate | Prostate | Micro-Ultrasound | Prostate | Dataset (participants row 17) |
| MMOTU | Pelvic Cavity | Ultrasound | Tumors | Dataset (participants row 34) |
| MuscleUS | Leg, Biceps | Ultrasound | Muscles | Dataset (participants row 31) |
| ps-fh-aop-2023 | Pelvis | Ultrasound | Pubic symphysis Fetal head | Dataset (organizers row 106) |
| RESECT | Brain | US | tumor | Dataset (organizers row 157) |
| SegThy | Neck | US | Thyroid, Carotid, Jugular Vein | Dataset (organizers row 126) |
| US simulation & segmentation | Abdomen | Ultrasound | Abdomen organs | Dataset (organizers row 111) |
| US-Nerve | Neck | US | Brachial Plexus (BP) | Dataset (organizers row 112) |
Chest radiography (4)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| Chest X-Ray Pneumothorax | Chest | CXR | Pneumothorax | Dataset (organizers row 63) |
| Chest Xray Masks and Labels | Chest | CXR | Lungs | Dataset (organizers row 64) |
| COVID-19 Radiography Database | Chest | CXR | Lungs | Dataset (organizers row 67) |
| COVID-QU-Ex | Chest | CXR | Lungs | Dataset (organizers row 68) |
Radiography (11)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| ARCADE | Chest | X-Ray | Vessel | Dataset (organizers row 52) |
| ASOCA | Vessel | X-Ray | Coronary Arteries | Dataset (organizers row 53) |
| BTXRD | Bone | X-Ray | Bone tumors | Dataset (organizers row 144) |
| DenPAR: Annotated Intra-Oral Periapical Radiographs Dataset for Machine Learning | Tooth | X-ray | Tooth and teeth | Dataset (participants row 77) |
| HipXRay | Pelvic Cavity | X-ray | Bones | Dataset (participants row 36) |
| PanDental | Dental | X-ray | Teeth,Jaw | Dataset (participants row 40) |
| PAXRay++ | Lung | X-ray | Thorax | Dataset (participants row 41) |
| PENGWIN | Pelvis | X-Ray | Pelvic fracture | Dataset (organizers row 161) |
| SCR (images from JSRT) | Lung | X-ray | Lungs, heart, clavicle | Dataset (participants row 38) |
| STS-2D-Tooth | Oral | X-Ray | tooth | Dataset (organizers row 142) |
| ToothSeg | Head and Neck | X-ray | Teeth | Dataset (participants row 39) |
Mammography (1)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| CDD-CESM | Chest | Mammography | lesion | Dataset (organizers row 61) |
PET and nuclear medicine (3)
Ophthalmic imaging (13)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| CHASE | Retinal | Fundus photography | Veins | Dataset (participants row 16) |
| COph100 | Eye | Fundus | Retinal vessel | Dataset (organizers row 145) |
| Drive | Vessel | Fundus | Fundus Blood vessel | Dataset (organizers row 71) |
| EOphtha | Retinal | Retinal | Eye microaneurysms | Dataset (participants row 29) |
| FairSeg | Eye | Fundus | Cup-Disc | Dataset (participants row 76) |
| FIVES | Eyes | Fundus | Vessel | Dataset (participants row 15) |
| IDRiD | Eye | Fundus | Lesions and Optic Disc | Dataset (organizers row 84) |
| Intraretinal Cystoid Fluid | Eye | OCT | Cystoid Macular Edema | Paper (organizers row 85) |
| OCT5k | retina | Optical Coherence Tomography | retinal layers | Dataset (organizers row 138) |
| OCTA500 | Head and Neck | OCT | Veins | Dataset (participants row 28) |
| PAPILA | Eye | Fundus | Optic Disc (OD) Optic Cup (OC) | Dataset (organizers row 102) |
| RAVIR | Eye | Fundus | Retinal Arteries and Veins | Dataset (organizers row 107) |
| STARE | Head and Neck | Fundus | Veins | Dataset (participants row 14) |
Pathology (8)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| Automated Gleason Grading Challenge 2022 | Pelvis | Pathology (H&E stained) | Prostate cancer | Dataset (organizers row 54) |
| Breast Cancer Segmentation | Cell | Pathology (H&E stained) | Breast cancer | Dataset (organizers row 58) |
| GlaS@MICCAI’2015 | Abdomen | Pathology (H&E stained) | Gland | Dataset (organizers row 78) |
| HuBMAP - Hacking the Kidney | Abdomen | Pathology (PAS Stained and H&E stained) | glomeruli FTU | Dataset (organizers row 81) |
| HuBMAP - HPA | Abdomen | Pathology | functional tissue unit (FTU) | Dataset (organizers row 82) |
| MonuSeg | Chest | Pathology (H&E stained) | tumor cell nuclei | Dataset (organizers row 94) |
| PanNuke | Cell | Pathology (H&E stained) | nuclei | Dataset (organizers row 101) |
| Wsss4luad | Chest | Pathology (H&E stained) | Lung adenocarcinoma | Dataset (organizers row 118) |
Dermatology (3)
Endoscopy and surgical video (10)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| BKAI-IGH NeoPolyp | Abdomen | Endoscopy | neoplastic polyps non-neoplastic polyps | Dataset (organizers row 55) |
| CholecSeg8k | Abdomen | Endoscopy | Tissues and instruments | Dataset (organizers row 65) |
| DFU | Abdomen | Endoscopy | Liver | Dataset (organizers row 70) |
| Endoscapes2023 | Abdomen | Endoscopy | semantic segmentation (4 anatomical structures, 1 descriptor of surgical dissection, and 1 tool class) | Dataset (organizers row 136) |
| FetoPlac | Vessel | Fetoscopic optical camera | vessel | Dataset (organizers row 74) |
| Kvasir-SEG | Abdomen | Endoscopy | Polyp | Dataset (organizers row 87) |
| m2caiSeg | Abdomen | Endoscopy | Tissues | Dataset (organizers row 90) |
| P2ILF | Abdomen | Endoscopy | Liver | Dataset (organizers row 99) |
| SAR-RARP50 | Prostate Cancer | laporoscopy | Instruments | Dataset (participants row 72) |
| SUN-SEG | Abdomen | Endoscopy | Polyp | Dataset (organizers row 132) |
Microscopy (20)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| AxonEM | Cells | Electron microscopy | Axon | Dataset (organizers row 152) |
| BBBC Collection (24,27,32,33,34,46,50) | Cells | Microscopy | Cells | Dataset (participants row 55) |
| BBC038 | Cells | Microscopy | Cells | Dataset (participants row 18) |
| CellTrackingChallenge (Fluo C3DH/N3DH) | Cells | Microscopy | Cell | Dataset (participants row 59) |
| CREMI | Drosophila brain | Electron microscopy | neuron | Dataset (organizers row 121) |
| Drosophila | Drosophila | Microscopy | Mitochondria, Synapses | Dataset (participants row 56) |
| ISBI_EM | Cells | Microscopy | neurons | Dataset (participants row 20) |
| LungVis1.0 | Lung | light sheet fluorescence microscopy | Airway | Dataset (organizers row 153) |
| MUCIC_Colon Tissue | Colon Tissue | virtual microscope | Cell nuclei | Dataset (organizers row 151) |
| MUCIC_Granulocytes | Cells | virtual microscopy | Cell nuclei | Dataset (organizers row 150) |
| MUCIC_HL60 | Cells | virtual microscopy | Cell nuclei | Dataset (organizers row 149) |
| NeurIPS CellSeg | Cell | Microscopy | cells | Dataset (organizers row 97) |
| NIS3D | Cells | Light Sheet Microscopy | nuclei | Dataset (organizers row 154) |
| NucMM-Z | zebrafish brain | Electron microscopy | nuclei | Dataset (organizers row 122) |
| PCMMD | Cells | Microscopy | plasma cells, non-plasma cell | Dataset (organizers row 141) |
| SELMA3D | Brain | Light-sheet Microscopy | neural activity | Dataset (organizers row 123) |
| SLiMIA | Cell | Microscopy | spheroid | Dataset (organizers row 137) |
| ssTEM | Cells | Microscopy | cells | Dataset (participants row 19) |
| UroCell | Urinary bladders | Electron microscopy | Endolysosomes and mitochondria | Dataset (organizers row 119) |
| WBC | Cells | Microscopy | Nuceleus, Cytoplasm | Dataset (participants row 21) |
Multimodal (5)
| Dataset | Anatomy | Modality as listed | Segmentation target | Source |
|---|---|---|---|---|
| HaN-Seg | Head and Neck | CT + MR | Head | Dataset (organizers row 9) |
| MedSeg | All | CT + MR | Vasulature, Muscles, Regions | Dataset (participants row 70) |
| MM-WHS | Chest | CT, MR | Heart | Dataset (organizers row 92) |
| Qubiq | Abdomen | CT,MR | Brain, kidney, pancreas, prostate | Dataset (participants row 11) |
| TRUSTED | — | CT + ultrasound | — | Dataset (organizers row 162) |
Before choosing a dataset
Write down the target population, acquisition, task, and clinical unit of evaluation. Then check whether the collection actually supplies the required inputs and reference labels. A challenge release may contain preprocessed volumes that conceal the work needed to accept routine hospital data. A selected-image dataset may exclude the technical failures that dominate a live service.
Record the exact release, download date, license or agreement, file checksums, label dictionary, split manifest, and preprocessing. Keep patient and source-examination identifiers in a protected linkage system when permitted, while using appropriate research identifiers in analysis. Never assume two differently named datasets are independent; derivatives and repeated releases can overlap.
| Question | What to record |
|---|---|
| Who is represented? | Sites, inclusion criteria, population, exclusions, missingness |
| What was acquired? | Modality, device, protocol, image type, geometry, timing |
| What is the reference? | Annotators, adjudication, clinical outcome, information available to labelers |
| What is independent? | Patient, examination, procedure, specimen, and derivative relationships |
| What use is permitted? | Access agreement, redistribution, commercial use, retention obligations |
| What does the test set establish? | Target population, fixed operating point, external validity, uncertainty |
A minimal research manifest
Use one row per defined source unit, with separate linked tables if the hierarchy is complex. Useful fields include a research patient key, examination key, source file or object identifiers, acquisition type, label version, split assignment, and exclusion reason. Do not put identifying clinical information into a public manifest.
Patient-level splitting is a starting point, not a complete validation strategy. Site-held-out, temporal, device-held-out, or prospective evaluation may be needed for the intended claim. Keep the final test set untouched while selecting models and thresholds. Report technical exclusions with denominators instead of quietly removing them.
A dataset absent from this directory may still be useful. Add it only after documenting the release, task, access terms, and primary source. Avoid aggregate counts that mix incompatible editions or count correlated derivatives as independent cases.