[Rtk-users] Add noise to projections

[Simon Rit]

Hi,
There is an AdditiveGaussianNoiseImageFilter
<https://github.com/SimonRit/RTK/blob/master/include/rtkAdditiveGaussianNoiseImageFilter.h>.
In practice, I personally do it in numpy with the python wrappings, see this
article <https://doi.org/10.1118/1.4945418> where I used this code:

I0=1e7
dH2O=0.01879 #mm^-1 at 75 keV
if I0!=0:
    slarray = rtk.GetArrayFromImage(sl)
    slarray = I0*np.exp(-1.*dH2O*slarray)
                                                      slarray =
np.maximum(np.random.poisson(slarray), 1)
    slarray = np.log(I0/slarray)/dH2O
    slarray = rtk.GetImageFromArray(slarray.astype(np.float32))
    slarray.CopyInformation(sl)
    sl = slarray

and wrote "The same simulations were repeated with Poisson noise.
The Shepp Logan densities were weighted by 0.01879 mm −1 ,
i.e., the linear attenuation coefficient of water at 75 keV. The
number of photons received per detector pixel without object
in the beam was constant for all pixels in both geometries and
equal to 10^7 ."
Simon


On Fri, May 15, 2020 at 10:04 AM <gabriele.belotti.bergamo at gmail.com> wrote:

> Dear RTK users and developers,
>
> I’m creating a CBCT dataset from existing CTs and would like to introduce
> noise into the generated projections before reconstruction.
> Is there an RTK implementation for this feature? Or is there any software,
> to you knowledge, that would efficiently add this noise (in terms of
> computation time)?
>
>
>
> Thanks in advance,
> Gabriele Belotti
> _______________________________________________
> Rtk-users mailing list
> Rtk-users at public.kitware.com
> https://public.kitware.com/mailman/listinfo/rtk-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://public.kitware.com/pipermail/rtk-users/attachments/20200515/9770c931/attachment.html>