[ITK] Choose the number of decimal

[Bradley Lowekamp]

Hello,

Depending on what exactly you want to do and the interpretation of the value, you could multiply the image by 100 and then convert to an appropriate integer representation to emulate fixed point operations for certain image processing algorithms. This poor mans fixed point representation may be trouble-some and cause confusion and errors if not carefully managed. However, I would think that just performing the rounding at the end of the computation would be sufficient for many cases,

Brad

On Dec 2, 2014, at 3:55 AM, Cyril Jaudet <drcjaudet at gmail.com> wrote:

> Hello everyone,
>  
> I have a simple yet a little stupid question. I want to restrain to two the number of decimal in float type image (PET SUV image). For this type of image it is really enought and will reduce the amount of time to postreatment, i suppose. An approximate value will be better then truncate. 
>  
> Is there a filter in itk or simple itk to do it?
>  
> Thank's a lot,
> Cyril
> _______________________________________________
> Community mailing list
> Community at itk.org
> http://public.kitware.com/mailman/listinfo/community