[Neuroimaging] regarding other denoising tecniques for 3d mri datasets

[Samuel St-Jean]

I don't see why it would not be possible, of course you would to roll out
your own version.

I have to admit I do not fully understand the question, so here is what I
could make out of it :

For the lmmse approach, there is a dwi version made for denoising and the
authors give some matlab code [1], so that would give you a starting point
to turn it into a 3D version (I don't remember all the inner details, so
maybe a 3D formulation is not valid, but anyway, that is the version I know
of, the statistical framework itself is probably valid for any type of
data).

As for the lib you mention, if it is this one
https://github.com/stefanv/supreme, running the setup.py would do the job.
Or you could try writing a nice email to the author also (he might even be
reading this mailing list, who knows) for more specific answers.

Thirdly, I also remember you asked about piesno, as a word of caution, it
is designed as an automatic noise estimator for repeated 3D volumes (dmri,
fmri, anything kind of time series related I'd say) relying on the
'flatness' of the signal along the 4th dimension. So, it would
unfortunately not work in theory for 3D mri volumes. I am not saying it is
impossible, just to be careful on how you feed the data to the function. In
any case, you can have a look at the inner piesno function as the public
one is a thin wrapper over slices, so it is possible to modify the
iterations to work over 3D smartly provided you segment the background
yourself for example or even just use the noise estimator directly over
segmented noise, bypassing the histogram estimation.

2016-06-09 14:44 GMT+02:00 Vivek Joshi <vivekjoshi1894 at gmail.com>:

> Hello
> The denoising technique mentioned in the dipy examples is NLM means
> technique.
> Is it possible to denoise the Datasets with 3d wavelet subband mixing
> tecnique and a LMMSE statistical approach? In case of 3d wavelet subband
> mixing, how to install supreme.lib module so that we can access pywt?
> Thank you
>
> _______________________________________________
> Neuroimaging mailing list
> Neuroimaging at python.org
> https://mail.python.org/mailman/listinfo/neuroimaging
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/neuroimaging/attachments/20160610/adedbe4b/attachment.html>