2D MRI liver slices with navigator frames. A training and test data set for image based 4D MRI reconstruction. (Part II)
No Thumbnail Available
Date
2021
Journal Title
Journal ISSN
Volume Title
Publisher
Otto-von-Guericke-Universität Magdeburg
Abstract
This data contains fully anonymized MRI data of the abdomens of 20 healthy subjects as basis to reconstruct 4D MRIs of the breathing motion of the subjects’ abdomen. This data publication uses the same protocol and is a continuation (part 2) of the data publication “2D MRI liver slices with navigator frames. A test data set for image based 4D MRI reconstruction” (https://doi.org/10.24352/UB.OVGU-2019-093)
Description
This data contains fully anonymized MRI data of the abdomens of 20 healthy subjects as basis to reconstruct 4D MRIs of the breathing motion of the subjects’ abdomen. This data publication uses the same protocol and is a continuation (part 2) of the data publication “2D MRI liver slices with navigator frames. A test data set for image based 4D MRI reconstruction” (https://doi.org/10.24352/UB.OVGU-2019-093)
The data is organized in three layers of directories.
The first layer of directories follows the naming convention YYYY-MM-DD_PPPP.zip, denoting the day of the session (YYYY-MM-DD) and an anonymized ID (PPPP) of the imaged subject.
Inside each zip there are four folders:
- Nav_+_Data -> contains the interleaved sequences.
- Nav_Pur_1 -> both Nav_Pur folders contain a reference navigator sequence.
- Nav_Pur_2 -> both Nav_Pur folders contain a reference navigator sequence.
- StarVibe_1 -> contains slices of a 3D Volume from the subjects abdomen
Some subjects contain only the star vibe data, as the other data is published in Part one already.
Use the PPPP identifierer to identify the same subject in all parts of the data publication
- Part one: "2D MRI liver slices with navigator frames. A test data set for image based 4D MRI reconstruction" (https://doi.org/10.1371/journal.pone.0235175)
- Part two: "2D MRI liver slices with navigator frames. a test data set for image based 4D MRI reconstruction. Part II" (https://doi.org/10.24352/UB.OVGU-2021-071)
All DICOM file names contain the following parts:
PPPP_XXX_XXXX.MR.GULAMHUSSENE_LEBER.SSSS.NNNN.YYYY.MM.DD
PPPP is an anonymize subject ID (unique and shared across data sets like Part I of this data set)
MR.GULAMHUSSENE_LEBER is the name of the study
SSSS is the sequence number
NNNN is the image number in the sequence (ATTANTION: this is not the acquisition order, use the DICOM tag AcquisitionTime to infer the acquisition order)
YYYY.MM.DD is the date of acquisition
The interleaved sequences are an alternating succession of navigator and data frames. Each data frame was imaged 166ms after the previous and 166ms before the following navigator frame.
Each interleaved sequence images another data slice.
The navigator frames in the interleaved as well as in the reference sequences are imaged in the same slice position (scanner coordinates).
The data is organized in three layers of directories.
The first layer of directories follows the naming convention YYYY-MM-DD_PPPP.zip, denoting the day of the session (YYYY-MM-DD) and an anonymized ID (PPPP) of the imaged subject.
Inside each zip there are four folders:
- Nav_+_Data -> contains the interleaved sequences.
- Nav_Pur_1 -> both Nav_Pur folders contain a reference navigator sequence.
- Nav_Pur_2 -> both Nav_Pur folders contain a reference navigator sequence.
- StarVibe_1 -> contains slices of a 3D Volume from the subjects abdomen
Some subjects contain only the star vibe data, as the other data is published in Part one already.
Use the PPPP identifierer to identify the same subject in all parts of the data publication
- Part one: "2D MRI liver slices with navigator frames. A test data set for image based 4D MRI reconstruction" (https://doi.org/10.1371/journal.pone.0235175)
- Part two: "2D MRI liver slices with navigator frames. a test data set for image based 4D MRI reconstruction. Part II" (https://doi.org/10.24352/UB.OVGU-2021-071)
All DICOM file names contain the following parts:
PPPP_XXX_XXXX.MR.GULAMHUSSENE_LEBER.SSSS.NNNN.YYYY.MM.DD
PPPP is an anonymize subject ID (unique and shared across data sets like Part I of this data set)
MR.GULAMHUSSENE_LEBER is the name of the study
SSSS is the sequence number
NNNN is the image number in the sequence (ATTANTION: this is not the acquisition order, use the DICOM tag AcquisitionTime to infer the acquisition order)
YYYY.MM.DD is the date of acquisition
The interleaved sequences are an alternating succession of navigator and data frames. Each data frame was imaged 166ms after the previous and 166ms before the following navigator frame.
Each interleaved sequence images another data slice.
The navigator frames in the interleaved as well as in the reference sequences are imaged in the same slice position (scanner coordinates).
Keywords
Training data, 4D-MRI, self-gated, respiratory motion, retrospective stacking, deep learning
Citation
TBA