Skip to content

Latest commit

 

History

History
154 lines (120 loc) · 10 KB

L2R-OASIS.md

File metadata and controls

154 lines (120 loc) · 10 KB

L2R-OASIS

Dataset Information

The Open Access Series of Imaging Studies (OASIS) is a project aimed at providing the scientific community with free access to neuroimaging datasets for the purpose of advancing discoveries in basic and clinical neuroscience. There are four versions available on the OASIS official website, and this description refers to the OASIS dataset version used in the L2R Challenge (Medical Image Registration in MELBA). This dataset is used for segmentation and registration and contains a total of 416 3D MR datasets, with label masks for segmentation and masks for registration.

The dataset consists of 35 label classes, which are as follows: 1 left cerebral white matter, 2 left cerebral cortex, 3 left lateral ventricle, 4 left inferior lateral ventricle, 5 left cerebellar white matter, 6 left cerebellar cortex, 7 left thalamus, 8 left caudate, 9 left putamen, 10 left pallidum, 11 third ventricle, 12 fourth ventricle, 13 brain stem, 14 left hippocampus, 15 left amygdala, 16 left accumbens area, 17 left ventral DC, 18 left vessels, 19 left choroid plexus, 20 right cerebral white matter, 21 right cerebral cortex, 22 right lateral ventricle, 23 right inferior lateral ventricle, 24 right cerebellar white matter, 25 right cerebellar cortex, 26 right thalamus, 27 right caudate, 28 right putamen, 29 right pallidum, 30 right hippocampus, 31 right amygdala, 32 right accumbens area, 33 right ventral DC, 34 right vessels, 35 right choroid plexus. Hosted by central.xnat.org, the OASIS dataset provides the community with an important open-access platform of neuroimaging data and post-processed imaging data covering a broad demographic, cognitive, and genetic spectrum, which can be used for neuroimaging, clinical, and cognitive studies on normal aging and cognitive decline. All data is accessible via www.oasis-brains.org.

Dataset Meta Information

Dimensions Modality Task Type Anatomical Structures Anatomical Area Number of Categories Data Volume File Format
3D MR Registration, Segmentation Brain Head and Neck 35 416 .nii.gz

Resolution Details

Dataset Statistics spacing (mm) size
min (1, 1, 1) (160, 224, 192)
median (1, 1, 1) (160, 224, 192)
max (1, 1, 1) (160, 224, 192)

Number of 2D slices in the dataset: 79,488 (based on the actual count from the training set of 414 cases that we downloaded).

Label Information Statistics

Label Cases Percentage Max Volume (cm³) Min Volume (cm³) Median Volume (cm³)
left_cerebral_white_matter 414 100.00% 310.84 218.81 267.775
left_cerebral_cortex 414 100.00% 368.4 216.04 298.06
left_lateral_ventricle 414 100.00% 85.84 5.61 14.525
left_inferior_horn_of_the_lateral_ventricle 414 100.00% 6.98 0.06 0.45
left_cerebellar_white_matter 414 100.00% 22.23 11.34 15.815
left_cerebellar_cortex 414 100.00% 83.69 50.26 66.615
left_thalamus 414 100.00% 10.36 6.38 8.52
left_caudate_nucleus 414 100.00% 7.06 3.04 4.735
left_putamen 414 100.00% 9.72 4.16 6.93
left_globus_pallidus 414 100.00% 2.75 1.44 2.13
third_ventricle 414 100.00% 5.19 0.87 1.59
fourth_ventricle 414 100.00% 4.79 1.2 2.29
brainstem 414 100.00% 32.2 18.33 25.01
left_hippocampus 414 100.00% 7.36 2.88 5.375
left_amygdala 414 100.00% 2.57 0.92 2.01
left_septal_area 414 100.00% 1.31 0.48 0.9
left_ventricle 414 100.00% 5.9 2.62 4.86
left_vessels 413 99.76% 0.43 0.0 0.06
left_choroid_plexus 414 100.00% 3.22 0.35 1.01
right_cerebral_white_matter 414 100.00% 320.36 223.78 269.05
right_cerebral_cortex 414 100.00% 360.81 208.48 296.715
right_lateral_ventricle 414 100.00% 78.74 5.14 13.1
right_inferior_horn_of_the_lateral_ventricle 414 100.00% 7.29 0.1 0.51
right_cerebellar_white_matter 414 100.00% 21.26 10.99 15.565
right_cerebellar_cortex 414 100.00% 84.79 49.67 66.265
right_thalamus 414 100.00% 10.04 6.28 8.34
right_caudate_nucleus 414 100.00% 7.4 3.21 4.96
right_putamen 414 100.00% 9.29 4.13 6.895
right_globus_pallidus 414 100.00% 2.74 1.32 2.13
right_hippocampus 414 100.00% 7.11 2.32 5.445
right_amygdala 414 100.00% 2.9 1.01 2.21
right_septal_area 414 100.00% 1.19 0.41 0.81
right_ventricle 414 100.00% 5.63 2.74 4.715
right_vessels 411 99.28% 1.53 0.0 0.05
right_choroid_plexus 414 100.00% 3.65 0.46 1.31

Visualization

Segmentation data.

Registration mask data.

File Structure

In the dataset, imagesTr are the training set brain MR files, imagesTs are the test set brain MR files, labelsTr are the training set labels, and maskTr and maskTs are the mask files used for registration in the test set.

Dataset
├── changelog.txt
├── OASIS_dataset.json
├── imagesTr
│   ├── OASIS_0001_0000.nii.gz
│   ├── OASIS_0002_0000.nii.gz
│   ├── ...
├── imagesTs
│   ├── OASIS_0415_0000.nii.gz
│   ├── OASIS_0416_0000.nii.gz
│   ├── ...
├── labelsTr
│   ├── OASIS_0001_0000.nii.gz
│   ├── OASIS_0002_0000.nii.gz
│   ├── ...
├── masksTr
│   ├── OASIS_0001_0000.nii.gz
│   ├── OASIS_0002_0000.nii.gz
│   ├── ...
├── masksTr
│   ├── OASIS_0415_0000.nii.gz
│   ├── OASIS_0416_0000.nii.gz
│   ├── ...
└── README.txt

Authors and Institutions

Daniel S. Marcus (Washington University)

Tracy H. Wang (Washington University)

Jamie Parker (Harvard University)

John G. Csernansky (Washington University)

John C. Morris (Washington University)

Randy L. Buckner (Harvard University, Massachusetts General Hospital, Howard Hughes Medical Institute)

Source Information

Official Website: https://learn2reg.grand-challenge.org/Datasets/

Download Link: https://www.oasis-brains.org/#data

Article Address: https://dash.harvard.edu/bitstream/handle/1/33896768/Buckner_OpenAccess.pdf?sequence=2

Publication Date: 2007

Citation

@article{marcus2007open,
  title={Open Access Series of Imaging Studies (OASIS): cross-sectional MRI data in young, middle aged, nondemented, and demented older adults},
  author={Marcus, Daniel S and Wang, Tracy H and Parker, Jamie and Csernansky, John G and Morris, John C and Buckner, Randy L},
  journal={Journal of cognitive neuroscience},
  volume={19},
  number={9},
  pages={1498--1507},
  year={2007},
  publisher={MIT Press One Rogers Street, Cambridge, MA 02142-1209, USA journals-info~…}
}

Original introduction article is here.