MedIMeta: A comprehensive and easy-to-use multi-domain multi-task medical imaging meta-dataset

We introduce the Medical Imaging Meta-Dataset (MedIMeta), a novel multi-domain, multi-task meta-dataset designed to facilitate the development and standardised evaluation of ML models and cross-domain few-shot learning algorithms for medical image classification. MedIMeta contains 19 medical imaging datasets spanning 10 different domains and encompassing 54 distinct medical tasks, offering opportunities for both single-task and multi-task training. All tasks are standardised to the same format and readily usable in PyTorch or other ML frameworks. All datasets have been previously published with an open license that allows redistribution or we obtained an explicit permission to do so.

Each dataset within the MedIMeta dataset is standardized to a size of 224 × 224 pixels which matches image size commonly used in pre-trained models. Furthermore, the dataset comes with pre-made splits to ensure ease of use and standardized benchmarking. We release a user-friendly Python package to directly load images for use in PyTorch.

Links

Project website: https://www.woerner.eu/projects/medimeta/

Data loading code (medimeta Python package): https://github.com/StefanoWoerner/medimeta-pytorch

Data creation code: https://github.com/StefanoWoerner/medimeta-dataset-scripts

Dataset Overview

Dataset Name
Dataset ID
License
Domain
Task Names
Task Targets
# Labels

AML Cytomorphology
aml
CC BY-SA 4.0
Microscopy
morphological class
multi-class classification
15

Breast Ultrasound
bus
CC BY-SA 4.0
Breast ultrasound
case categorymalignancy
multi-class classificationbinary classification
32

Colorectal Cancer Histopathology
crc
CC BY-SA 4.0
Histopathology
tissue class
multi-class classification
9

Chest X-ray Multi-disease
cxr
CC BY-SA 4.0
Chest X-ray
disease labelspatient sex
multi-label classificationbinary classification
142

Dermatoscopy
derm
CC BY-SA 4.0
Dermatoscopy
disease category
multi-class classification
7

Diabetic Retinopathy (Regular Fundus)
dr_regular
CC BY-SA 4.0
Retinal fundus
DR levelOverall qualityArtifactClarityField definition
ordinal regressionbinary classificationordinal regressionordinal regressionordinal regression
52655

Diabetic Retinopathy (Ultra-widefield Fundus)
dr_uwf
CC BY-SA 4.0
Retinal fundus
DR level
ordinal regression
5

Fundus Multi-disease
fundus
CC BY-SA 4.0
Retinal fundus
disease presencedisease labels
binary classificationmulti-label classification
245

Glaucoma-specific fundus images
glaucoma
CC BY-SA 4.0
Retinal fundus
Glaucoma suspect
binary classification
2

Mammography (Calcifications)
mammo_calc
CC BY-SA 4.0
Mammography
pathologycalc typecalc distribution
binary classificationmulti-label classificationmulti-label classification
2145

Mammography (Masses)
mammo_mass
CC BY-SA 4.0
Mammography
pathologymass shapemass margins
binary classificationmulti-label classificationmulti-label classification
285

OCT
oct
CC BY-SA 4.0
OCT
disease classurgent referral
multi-class classificationbinary classification
42

Axial Organ Slices
organs_axial
CC BY-NC-SA 4.0
Abdominal CT
organ label
multi-class classification
11

Coronal Organ Slices
organs_coronal
CC BY-NC-SA 4.0
Abdominal CT
organ label
multi-class classification
11

Sagittal Organ Slices
organs_sagittal
CC BY-NC-SA 4.0
Abdominal CT
organ label
multi-class classification
11

Peripheral Blood Cells
pbc
CC BY-SA 4.0
Microscopy
cell class
multi-class classification
8

Pediatric Pneumonia
pneumonia
CC BY-SA 4.0
Chest X-ray
pneumonia presencedisease class
binary classificationmulti-class classification
23

Skin Lesion Evaluation (Dermoscopy)
skinl_derm
CC BY-SA 4.0
Dermatoscopy
DiagnosisDiagnosis groupedPigment NetworkBlue Whitish VeilVascular StructuresVascular Structures groupedPigmentationPigmentation groupedStreaksDots and GlobulesRegression StructuresRegression Structures grouped
multi-class classificationmulti-class classificationmulti-class classificationbinary classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationbinary classification
1553283533342

Skin Lesion Evaluation (Clinical Photography)
skinl_photo
CC BY-SA 4.0
Clinical skin imaging
DiagnosisDiagnosis groupedPigment NetworkBlue Whitish VeilVascular StructuresVascular Structures groupedPigmentationPigmentation groupedStreaksDots and GlobulesRegression StructuresRegression Structures grouped
multi-class classificationmulti-class classificationmulti-class classificationbinary classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationmulti-class classificationbinary classification
1553283533342