Closed gb4944 closed 6 years ago
Please test the latest commit and close this issue if it fixes your issues. Note that PAR/REC is deprecated by Philips. In general, dcm2niix seems to fast and relatively robust solution. The one limitation is if you abort a 4D scan part way through and the number of 2D slices are not evenly divisible by the number of 3D volumes. In this case, dcm2niix will not convert the file, but it will tell you that you can salvage your data with dicm2nii.
With regards to specifying filenames (e.g. "-f %p_%s"), here are the conversions between the arguments and the PAR tags. Note that not all arguments have an equivalent (e.g. %i %j).
That fixed it, thank you very much!
Hi, I've encountered an issue with the file naming/JSON generation from PARREC, using the latest pre-release (v1.0.20180328). An incorrect string is inserted into file names and and the JSON
ProtocolName
tag. A minimal command likedcm2niix input_dir/Subcortex_0016_014_R01_SB_INV1_12_1.PAR
produces this type of file name:PARREC_Subcortex_0016_014_R01_SB_INV1======tool4.2_19770703150928_12.nii
The string
19770703150928
doesn't seem to appear anywhere in the .PAR file, whereas======tool4.2
seems to come from the first few lines of the file. The latter string is identical for all files I've tried, but the number series is different for each subject, and identical for different files for one subject.Modifying the file name removes the number series, but
======tool4.2
seems to be read as part of the %p tag.dcm2niix -o output_dir -b y -ba y -f "sub-${subject_id}_echo-%e_inv-%p" -z y ${input_file}
creates a file namedsub-003_echo-1_inv-SB_INV1======tool4.2.nii.gz
. The corresponding JSON contains the same string:"ProtocolName": "SB_INV1======tool4.2"
.Thank you