PhilipsDataList

This package contains basic functionally to read in *.{data,list} files exported using Philips' MR systems.

Installation

Activate the environment where you want to use this package, enter Pkg mode and add the package with the following command:

add git@gitlab.op.umcutrecht.nl:computational-imaging-lab/philipsdatalist.git

Package Functionality

read_data_list(path::String)

• This function reads samples from the .data file. There are a limited number of complex data vector types (e.g. STD, NOI). For each type, the samples are stored separately in a Vector{ComplexF32}. The vectors with samples are stored together in a NamedTuple with fieldnames corresponding to the different types (e.g. STD, NOI).
• The general scan information is extracted from the .list file and stored as info::Vector{String}.
• The attributes of each of the complex data vectors is extracted from the .list file. For each complex data vector type the attributes are stored in a DataFrame and the DataFrames are stored in a NamedTuple.
• The function returns samples_per_type, attributes_per_type and info.

This package only really reads the .{data,list} files and it does not process (e.g. sort) the data.

data_list_to_kspace(path::String)

• This function reads in the .{data,list} files using read_data_list and then it sorts the samples of type STD into a k-space.
• The k-space has named dimensions. The dimension names are found in DIMENSIONS_STD. By default, dimensions of size 1 are dropped.
• The kspace array is also an OffsetArray. This allows, for example, the k-space center(s) to be extracted as kspace[kx=0, ky=0, kz=0]. After FFT-ing, it make sense to get rid of the offsets.
• For highly undersampled scans, this approach is memory-inefficient because non-sampled readouts will be stored as zeros.

Background

The .data file is a binary file that contain measured samples in (complex) single precision floating point format.

For further processing of the data (e.g. sorting), note that the samples are grouped as complex data vectors (Philips' phrasing). There are several complex data vector types:

STD = Standard data vector (image data or spectroscopy data)
REJ = Rejected standard data vector
    (only for scans with arrhythmia rejection)
PHX = Correction data vector for EPI/GraSE phase correction
FRX = Correction data vector for frequency spectrum correction
NOI = Preparation phase data vector for noise determination
NAV = Phase navigator data vector
DNA = Dynamic phase navigator vector

Besides the type, there are several other attributes that complex data vectors have:

mix      = mixed sequence number
dyn      = dynamic scan number
card     = cardiac phase number
echo     = echo number
loca     = location number
chan     = synco channel number
extr1    = extra attribute 1 (semantics depend on type of scan)
extr2    = extra attribute 2 (   ''       ''   ''  ''  ''  '' )
ky,kz    = k-space coordinates in 1st and 2nd preparation direction (image data)
kx,ky,kz = k-space coordinates in 1st, 2nd and 3rd preparation direction (spectroscopy data)
aver     = sequence number of this signal average
sign     = sign of measurement gradient used for this data vector (1 = positive, -1 = negative)
rf       = sequence number of this rf echo (only for TSE, TFE, GraSE)
grad     = sequence number of this gradient echo (only for EPI/GraSE)
enc      = encoding time (only for EPI/GraSE)
rtop     = R-top offset in ms
rr       = RR-interval length in ms
size   = data vector size   in number of bytes (1 complex element = 2 floats = 8 bytes)
offset = data vector offset in number of bytes (first data vector starts at offset 0)

The .list file contains the attributes for each of the complex data vectors and which samples belong to which complex data vector (using size and offset). It also contains some general information of the scan such as the k-space and image space coordinate ranges, the number of locations, dynamics and averages.

Todo

• Store info in a DataFrame as well.
• Make a faster version in case all the STD samples are contiguous.