amid 0.7.0

A curated list of medical imaging datasets with unified interfaces

Awesome Medical Imaging Datasets (AMID) - a curated list of medical imaging datasets with unified interfaces

Getting started

Just import a dataset and start using it!

Note that for some datasets you must manually download the raw files first.

from amid.verse import VerSe

ds = VerSe()
# get the available ids
print(len(ds.ids))
i = ds.ids[0]

# use the available methods:
#   load the image and vertebrae masks
x, y = ds.image(i), ds.masks(i)
print(ds.split(i), ds.patient(i))

# or get a namedTuple-like object:
entry = ds(i)
x, y = entry.image, entry.masks
print(entry.split, entry.patient)

Available datasets

Name	Entries	Body region	Modality
AMOS	600	Abdomen	CT, MRI
BIMCVCovid19	16335	Chest	CT
CC359	359	Head	MRI T1
CT_ICH	75	Head	CT
CrossMoDA	484	Head	MRI T1c, MRI T2hr
EGD	774	Head	FLAIR, MRI T1, MRI T1GD, MRI T2
FLARE2022	2100	Abdomen	CT
LIDC	1018	Chest	CT
LiTS	201	Abdominal	CT
LiverMedseg	50	Chest, Abdomen	CT
MIDRC	155	Thorax	CT
MOOD	1358	Head, Abdominal	MRI, CT
Medseg9	9	Chest	CT
MoscowCancer500	979	Thorax	CT
MoscowCovid1110	1110	Thorax	CT
NLST	13623	Thorax	CT
NSCLC	422	Thorax	CT
RSNABreastCancer	54710	Thorax	MG
Totalsegmentator	1204	Head, Thorax, Abdomen, Pelvis, Legs	CT
VSSEG	242	Head	MRI T1c, MRI T2
VerSe	374	Thorax, Abdomen	CT

Check out our docs for a more detailed list of available datasets and their fields.

Install

Just get it from PyPi:

pip install amid

Or if you want to use version control features:

git clone https://github.com/neuro-ml/amid.git
cd amid && pip install -e .

Contribute

Check our contribution guide if you want to add a new dataset to AMID.