On Wed, May 11, 2011 at 03:22:19PM -0400, Panagiotis Foteinos wrote:
> I see that Gwyddion supports .hdr reading but no .hdr writing. Is it
> difficult to write an .hdr exporter? hdr is supported by ITK.
I don't think this is the same .hdr. AFAIK the only .hdr files Gwyddion
reads are the headers of Unisoku multi-part data.
> ITK also
> supports the creation of such images. So you might not even have to study
> the specs. DICCOM and NRRD formats are also very popular in medical
> community for data representation, since the pixel values can be literally
> anything (doubles, vectors, you name it). In fact, the output of CT and MRI
> scanners is DICOM series. Of course, the best choice is ITK's native image
> format MHA.
I looked at DICOM before but I got lost. It is thousands of pages of
convoluted specification legalese irrelevant to Gwyddion that does not
assign any medical meaning to the data it works with. In fact, Gwyddion
rarely cares about any meaning at all – it just offers the tools and the
user decides what makes sense (this is quite different from the medical
application where everyone needs to cover his ass). Is any example or
tutorial available to how to write a simple DICOM-compatible 2D data
file with no medical meaning implied?
I did not know NRRD; but it looks much saner and I will probably
implement it.
If MHA is the same MHA as shown here
http://code.google.com/p/brainseg/wiki/FileFormats
then this is also reasonable although I cannot find any proper MHA
specification. Apparently everyone assumes you use ITK to work with MHA
files... (Well, we are guilty of the same with GWY that has never been
meant to be an exchange format – but it is still specified at least on
the self-describing level.)
> Meanwhile, I might need to write my own "dummy" MHA exporter just to get the
> job done, since I have a timetable to follow. Which brings me to the next
> question: Is it better to start from the raw data (usually in excell format)
> and export to an ITK format, or start from the STP image? Precision-wise.
Raw data in Excel format can mean various things but, in principle, it
can represent double-precision floating point data.
I know at least two STP formats, the old Molecular Imaging PicoScan one
and the WSxM one. The former contains 16bit integral data, the latter
can contain different data, including double-precision floating point.
Using the raw data of whatever kind means that you won't lose precision
in the process by definition (except possibly in the final export step).
Going through other representations may lose precision but it depends.
> And my last (newbie-ish question) is as follows: I am searching for the
> specs of the STP format, but completely unsuccessfully... Does anybody know
> where to get it from?
Which one? I can probably find something about the PicoScan format, the
WSxM format description can be found on-line.
Regards,
Yeti
------------------------------------------------------------------------------
Achieve unprecedented app performance and reliability
What every C/C++ and Fortran developer should know.
Learn how Intel has extended the reach of its next-generation tools
to help boost performance applications - inlcuding clusters.
http://p.sf.net/sfu/intel-dev2devmay
_______________________________________________
Gwyddion-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/gwyddion-users
