neuropoly / data-management

Repo that deals with datalad aspects for internal use
4 stars 0 forks source link

New dataset `lumbar-rsna-challenge2024` #332

Open Nilser3 opened 1 month ago

Nilser3 commented 1 month ago

Description

I would like to push a new dataset lumbar-rsna-challenge2024 to git-annex. This is a public dataset, contains 1975 subjects from RSNA 2024 Lumbar Spine Degenerative Classification challenge : https://www.kaggle.com/competitions/rsna-2024-lumbar-spine-degenerative-classification

Details

Raw data (Dicom) location: duke/mri/lumbar-classification-rsna-challenge-2024 Project repo: Link Code for convert Dicom to BIDS : Link

Content

anat : T1w, T2w_sag and T2w_ax GT : annotation of the pathology (binary point). ie: spinal canal stenosis, neural foraminal narrowing and subarticular stenosis, from L1 to S1 region : L1 to Sacrum

Naming suggestions for the dataset

Following the intranet convention, maybe it could be: lumbar-rsna-challenge2024

feedback pls @simonqueric @NathanMolinier

Related issue

#14

mguaypaq commented 1 month ago

I created the dataset here, and gave write access to @Nilser3, @simonqueric, @NathanMolinier: https://data.neuro.polymtl.ca/datasets/lumbar-rsna-challenge-2024

Nilser3 commented 1 month ago

Thanks you @mguaypaq

@simonqueric the data is already in my branch: nlm/initial_data For each binary point (annotation) there is a Json file where the severity is indicated.

cat derivatives/labels/sub-1190473557/anat/sub-1190473557_acq-ax_rec-3634559093_T2w_desc-L1L2_label-LeftSubarticularStenosis_label.json

{
    "SpatialReference": "orig",
    "PathologySeverity": "Normal/Mild",
    "GeneratedBy": [
        {
            "Name": "Manual",
            "Author": "Nilser Laines Medina",
            "Date": "2024-07-22",
            "Note": "Based on original coordinates from sourcedata/train.csv in DICOM space"
        }
    ]
}
Nilser3 commented 1 month ago

@NathanMolinier @simonqueric @KaterinaKrejci231054 TotalSegmentations are git-annexed