[Neuroimaging] Reconstruction of the diffusion signal with the Tensor model after de-noising

[Samuel Pilgrim]

Ariel,
Sorry it took me a while to get back to you.  I just wanted to let you know
that your advice was quite helpful.  I believe I've more or less gotten
tensor fitting to work after applying the denoising procedure.  There are
still a number of other difficulties with the script that I'm working on,
so I'm looking forward to your visit to Franco's lab.

Thanks again,
Sam

On Fri, Oct 16, 2015 at 9:28 AM, Samuel Pilgrim <spilgrim at umail.iu.edu>
wrote:

> Ariel,
> Thanks a lot.  I'll try that out today/tomorrow and get back to you.
>
> Best,
> Sam
>
> On Fri, Oct 16, 2015 at 2:46 AM, Ariel Rokem <arokem at gmail.com> wrote:
>
>> Hi Sam,
>>
>> You might be able to draw inspiration from the new DKI example, in which
>> denoising is first applied:
>>
>> https://github.com/nipy/dipy/blob/master/doc/examples/reconst_dki.py#L124
>>
>> And then a model is fit (in this case the DKI model, but the TensorModel
>> would look very similar):
>>
>> https://github.com/nipy/dipy/blob/master/doc/examples/reconst_dki.py#L124
>> https://github.com/nipy/dipy/blob/master/doc/examples/reconst_dki.py#L145
>>
>> Does that work for you?
>>
>> Cheers,
>>
>> Ariel
>>
>>
>>
>> On Wed, Oct 14, 2015 at 2:21 PM, Samuel Pilgrim <spilgrim at umail.iu.edu>
>> wrote:
>>
>>> Neuroimaging mailing list,
>>>
>>> I'm an undergraduate experimenting with dipy and am currently working
>>> on a project that would combine a few of the examples given in the
>>> documentation.  In particular, I'd like to be able to take one of the
>>> sample data-sets, apply the de-noising procedure described in the
>>> documentation under "Denoise images using Non-Local Means" and then use
>>> the tensor model to reconstruct the diffusion signal using the de-noised
>>> data.  I have scripts that do each of these things separately, but when I
>>> try to do reconstruction with de-noised data, I get an alignment error on
>>> the line that reads "  tenfit = tenmodel.fit(maskdata)  ".  I believe
>>> the issue is caused by a command in the de-noising part of the script that
>>> goes  "  data = data[..., 1]  ".  After this line, the dimension of the
>>> object data has been reduced by one, (i.e. the command print(data.shape)
>>> returns a tuple of 3 integers instead of 4.  I think this is what's causing
>>> the alignment issue later, but I also can't get it to run without this
>>> line.  Any assistance would be greatly appreciated.
>>>
>>> Best,
>>> Sam
>>>
>>> _______________________________________________
>>> Neuroimaging mailing list
>>> Neuroimaging at python.org
>>> https://mail.python.org/mailman/listinfo/neuroimaging
>>>
>>>
>>
>> _______________________________________________
>> 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/20151028/839c9114/attachment.html>