The input of Data Factory is the output of Data Capture.
├── DICOM
│ └── 2016 -- yearly folder, date represents the date of export
│ └── 20161029 -- daily folder, date represents the date of export
│ └── scan_research_id -- see description below
│ └── dicom_name_generated_01.dcm -- set of DICOM files
│ └── dicom_name_generated_02.dcm -- set of DICOM files
│ └── dicom_name_generated_03.dcm -- set of DICOM files
└── EHR
└── 2016 -- yearly folder, date represents the date of export
└── 20161029 -- daily folder, date represents the date of export
├── table1.csv -- pre-defined name for 1st table containing EHR data, depends on hospital data
└── table2.csv -- pre-defined name for 2nd table containing EHR data, depends on hospital data
└── ... -- more (or less) tables as needed, depends on hospital data
├── NIFTI
│ └── 2016 -- yearly folder, date represents the date of export
│ └── 20161029 -- daily folder, date represents the date of export
│ └── scan_research_id -- see description below
│ └── dicom_name_generated_01.nifti -- Nifti file
│ └── dicom_name_generated_01.json -- metadata for the Nifti file
│ └── dicom_name_generated_02.nifti -- Nifti file
│ └── dicom_name_generated_02.json -- metadata for the Nifti file
└── EHR
└── 2016 -- yearly folder, date represents the date of export
└── 20161029 -- daily folder, date represents the date of export
├── table1.csv -- pre-defined name for 1st table containing EHR data, depends on hospital data
└── table2.csv -- pre-defined name for 2nd table containing EHR data, depends on hospital data
└── ... -- more (or less) tables as needed, depends on hospital data
In addition, currently our MRI image pre-processing pipelines are optimised to work only with well-defined medical images.