Table 2 Data summary of MedMNIST v2 dataset, including data source, data modality, type of the classification task together with the number of classes for multi-class or that of labels for multi-label, number of samples in total and in each data split (training/validation/test).

Name	Source	Data Modality	Task (# Classes/Labels)	# Samples	# Training/Validation/Test
MedMNIST2D
PathMNIST	Kather et al.^16,17	Colon Pathology	MC (9)	107,180	89,996/10,004/7,180
ChestMNIST	Wang et al.¹⁸	Chest X-Ray	ML (14) BC (2)	112,120	78,468/11,219/22,433
DermaMNIST	Tschandl et al.^19,20, Codella et al.²¹	Dermatoscope	MC (7)	10,015	7,007/1,003/2,005
OCTMNIST	Kermany et al.^22,23	Retinal OCT	MC (4)	109,309	97,477/10,832/1,000
PneumoniaMNIST	Kermany et al.^22,23	Chest X-Ray	BC (2)	5,856	4,708/524/624
RetinaMNIST	DeepDRiD Team²⁴	Fundus Camera	OR (5)	1,600	1,080/120/400
BreastMNIST	Al-Dhabyani et al.²⁵	Breast Ultrasound	BC (2)	780	546/78/156
BloodMNIST	Acevedo et al.^26,27	Blood Cell Microscope	MC (8)	17,092	11,959/1,712/3,421
TissueMNIST	Ljosa et al.²⁹	Kidney Cortex Microscope	MC (8)	236,386	165,466/23,640/47,280
OrganAMNIST	Bilic et al.³⁰, Xu et al.³¹	Abdominal CT	MC (11)	58,850	34,581/6,491/17,778
OrganCMNIST	Bilic et al.³⁰, Xu et al.³¹	Abdominal CT	MC (11)	23,660	13,000/2,392/8,268
OrganSMNIST	Bilic et al.³⁰, Xu et al.³¹	Abdominal CT	MC (11)	25,221	13,940/2,452/8,829
MedMNIST3D
OrganMNIST3D	Bilic et al.³⁰, Xu et al.³¹	Abdominal CT	MC (11)	1,743	972/161/610
NoduleMNIST3D	Armato et al.³²	Chest CT	BC (2)	1,633	1,158/165/310
AdrenalMNIST3D	New	Shape from Abdominal CT	BC (2)	1,584	1,188/98/298
FractureMNIST3D	Jin et al.³³	Chest CT	MC (3)	1,370	1,027/103/240
VesselMNIST3D	Yang et al.³⁴	Shape from Brain MRA	BC (2)	1,909	1,335/192/382
SynapseMNIST3D	New	Electron Microscope	BC (2)	1,759	1,230/177/352

Upper: MedMNIST2D, 12 datasets of 2D images. Lower: MedMNIST3D, 6 datasets of 3D images. MC: Multi-Class. BC: Binary-Class. ML: Multi-Label. OR: Ordinal Regression.

Search