[Insight-users] Problem writing to Analyze file format

[Oleksandr Dzyubak]

Daniele,

Thanks a lot for pointing me out to those documents.
Now the things became a bit clear.

One question though.
Some ITK filters (distance map, for example) work with a type double.
If I save the result in the Analyze format and
try to see the output using the ImageJ (also from NIH) viewer,
it always shows some garbage but it is OK if the float type used.

Does it mean that I have to use a float type in those filters?

Thanks,

Alex

Daniele Domenichelli wrote:
> Oleksandr Dzyubak wrote:
>   
>> The only problem is that the Analyze75 header does not have a tag with
>> that info, right?
>>     
>
> No, analyze has a tag "orient" (1 byte) but it has only 6 valid values
>
>           0         transverse unflipped
>           1         coronal unflipped
>           2         sagittal unflipped
>           3         transverse flipped
>           4         coronal flipped
>           5         sagittal flipped
>
> Nifti uses quaternion for orientation. See nifti header if you are
> interested:
>
>      http://nifti.nimh.nih.gov/pub/dist/src/niftilib/nifti1.h
>
>
>
>   
>> The Analyze75 header size is 348 vs 352 for NifTI. Does it mean that
>> those extra bites can hold all the directional info in case of the NIfTI
>> format and well treated by the ITK filters?
>>     
>
> No bytes from 349 to 352 are used for extensions.
> Nifti header in not necessarily 352 bytes, it can contain extensions for
> example to contain dicom tags (see nifti header for details)
> Nifti stores informations about orientations of the image in bytes that
> were not used in Analyze
>
>
>   
>> The question is since the Analyze image format has "poor performance" in
>> a sense
>> of the directional info, how does it affect the final results after ITK
>> filters have been used
>> on those images? Do any "info corruptions/changes" squeeze in
>> automatically into
>> the image info tags afterwards since it might be implemented in a filter
>> by some defaults?
>>     
>
> I think that the only thing you should take care of when using analyze
> files is that you can't write them in any direction, so you may need to
> reorient the image.
>
> I'm not sure about this anyway, because I mostly use nifti, maybe
> someone else on the mailing list who uses analyze can tell you something
> more...
>
>
> Regards,
> Daniele
>