Philips LIST/DATA raw format

[jsgi...@gmail.com]

Hi,

We are planning a pipeline for processing SMS DSI data for a multi-vendor TBI project starting from raw data from each platform. For the Philips we've decided to start with LIST/DATA rather than SIN/LAB/RAW since the steps between are better done using the Philips algorithms. 

Do you know of any effort to do this - no sense reinventing if it's already done.

Thanks,

Joe

Research Associate

Division of Magnetic Resonance Research
Department of Radiology and Radiological Science

Johns Hopkins University School of Medicine

[Michael Hansen]

I don't know of a LIST/DATA -> ISMRMRD converter. I have limited experience with that format, but I think it should be very easy to write the converter since the data is already sort of preprocessed a bit as you point out. If you write a converter, I am happy to help in terms of how to get started and some other general guidance. You can contact me directly if you would like to have a chat on the phone about it. I would also encourage you to share the converted if you make one. 

--
You received this message because you are subscribed to the Google Groups "ISMRMRD" group.
To unsubscribe from this group and stop receiving emails from it, send an email to ismrmrd+u...@googlegroups.com.
To post to this group, send email to ism...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/ismrmrd/54fed6ee-c5f1-4e58-b752-91e239e8565a%40googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Joseph S. Gillen]

I've made progress on integrating code for LIST/DATA conversion into the existing philips_to_ismrmrd code used for SIN/LAB/RAW format. I have some questions:

Raw data in the LIST/DATA file contains multiple data "types":

STD - standard acq

REJ - rejected standard data - arrythmia rejection - doesn't seem to have a place. If it appears, I suppose it can be skipped and not copied into the h5 file but we’re not using any cardiac gating. 

PHX - phase correction - I am setting the ISMRMRD_ACQ_IS_PHASECORR_DATA flag. 

FRX - frequency spectrum correction - there does not appear to be an appropriate flag - what to use? New flag?

NOI - setting the ISMRMRD_ACQ_IS_NOISE_MEASUREMENT flag

NAV - setting the ISMRMRD_ACQ_IS_NAVIGATION_DATA flag

DNA - dynamic navigator - same as NAV??

In the SMS acquisitions I'm using for test, I see only NOI, FRX, PHX and STD.

NOI - one long acq for each coil

FRX - 2 acqs for each coil for each slice - one with sign +1, one sign -1

PHX - 2 acqs for each coil for each slice at first PE - one with sign +1, one sign -1

the FRX and PHX acqs are 1/2 size of STD acqs - appears they are not oversampled

Philips encodes ky and kz using signed integers with center k-space at zero and negative and positive values. The kspace_encoding_step_1 and _2 of the ISMRMRD_EncodingCounters struct are unsigned integers. So I am shifting the values up by the minimum.

It appears that we’ll really need multiple .xsl files to map dimensions like fMRI dynamics or DTI/DSI vector or ASL types to the repetitions dimension of the schema. Is this how it was envisioned to work?

Thanks,

Joe