Author: rick reynolds (---.nimh.nih.gov)
Date: 10-26-09 17:10
Hi Christina,
I have not heard of anyone trying to use NIfTI for spectroscopy
before (which means little, except that such things have not
been posted here or on the NITRC site).
Similar to FMRI, I guess such datasets must be slices, possibly
over space (so volumes) and time (or at least some set of volumes
or slices). Is that the case?
What are the typical attributes of such a dataset (grid size,
orientation, dimensions, data types, statistics, etc.)?
One possible difficulty is that the dimensions are stored as
signed shorts. I'm not sure how high-resolution your images
might be, but signed short dimension sizes would imply a limit
of ~32K in any direction (e.g. a 32767 x 32767 slice). Is that
sufficient?
---
NIfTI is a pretty simple format, and has gotten popular (meaning
there are many readers/writers out there, with free APIs in at
least C and python). So if it meets the needs of spectroscopy
datasets, and there isn't a popular alternative, it might be a
very good way to go.
- rick
|
|
