dipy/dipy

Diffusion MR Imaging in Python

Rob Reid

@captainnova

Hi, I followed the AFQ tract profile tutorial, https://dipy.org/documentation/1.0.0./examples_built/afq_tract_profiles/#example-afq-tract-profiles , and my final figure does not fully match the one on the page. My AF plot matches, but my CST plot has FA staying between 0.4 and 0.1 until it drops to 0 at node 100. The CST plot on the tutorial page is 0 in most places, which does not seem reasonable. Maybe it needs updating? I am using dipy 1.0.0. Thanks!

Ariel Rokem

@arokem

@captainnova : thanks for the heads up. I agree with you that something's wrong with the results presented on that webpage!

@captainnova : see nipy/dipy#1985

Rob Reid

@captainnova

@arokem : thanks for checking and starting #1985. I've updated it with my result.

nis02002

@nis02002

hi everyone, how can extra the x-y location of water from freewater DTI model? We can plot the data as image, but I also want to check the exact value of locations

AlexBadea

@portokalh

I would be grateful for your thoughts on what could be wrong when feeding labels in utils.connectivity.matrix: M, grouping = utils.connectivitymatrix(streamlines, labels.astype(int), affine_labels,
return_mapping=False,
mapping_as_streamlines=False)

File "/Users/alex/py3/lib/python3.7/site-packages/dipy/tracking/utils.py", line 157, in connectivity_matrix
raise ValueError("label_volume must be a 3d integer array with labelvolume must be a 3d integer array withnon-negative label values
labels.shape= (200, 400, 200) when labels.max() = 332 and labels.min() = 0. Thanks much!

tried adding 1 to the labels, no difference!

Serge Koudoro

@skoudoro

Hi @portokalh, can you give us the results of labels.ndim ? labells.dtype.kind?

Futhermore, it seems there is an inversion between affine_labels and labels in your function arguments. To make sure that is correct, be explicit on your function:

utils.connectivitymatrix(streamlines, label_volume=labels.astype(int), affine=affine_labels,
return_mapping=False,
mapping_as_streamlines=False)

AlexBadea

@portokalh

Thanks @skoudoro : labels.dtype.kind='i'; labels.ndim=3; running what you suggested takes me to ile "/Users/alex/py3/lib/python3.7/site-packages/dipy/tracking/utils.py", line 164, in <listcomp>
endpoints = [sl[0::len(sl)-1] for sl in streamlines]
ValueError: slice step cannot be zero

2 replies

nis02002

@nis02002

hi everyone , in order to use water free DTI, do we have to get at least three different b-vales/b-vectors?

Ariel Rokem

@arokem

@nis02002 : at least two non-zero b-values, if you intend to use the method implemented here: https://github.com/nipy/dipy/blob/master/dipy/reconst/fwdti.py

There are new developments in the pipeline to remove this constraint, but these methods are not quite ready for general use yet

Hopefully in an upcoming release

zlf123-marker

@zlf123-marker

Excuse me, I want to know how to display t1 data and TRK files with dipy library?

Ariel Rokem

@arokem

@zlf123-marker : you should take a look at the examples on the website: https://dipy.org/tutorials/

zlf123-marker

@zlf123-marker

I have a look, but I want to display my TRK file and t1 data, I don't know what to change?Can you help me @arokem

robertl1996

@robertl1996

Does anybody know where the source code for this denoise library can be found: https://dipy.org/documentation/1.0.0./reference/dipy.denoise/. Particularly the nl means filters

Brian

@bhsilverstein

Hey folks - newbie question here: is there a simple way to concatenate data from two or more stateful tractograms into a single tractogram? I could just pull the streamlines and concatenate them, but I'd like all the metadata to come along with the streamlines.

Jon Haitz Legarreta Gorroño

@jhlegarreta

@robertl1996

Does anybody know where the source code for this denoise library can be found: https://dipy.org/documentation/1.0.0./reference/dipy.denoise/. Particularly the nl means filters

The source code is found here: https://github.com/nipy/dipy/blob/master/dipy/denoise/

Jon Haitz Legarreta Gorroño

@jhlegarreta

@zlf123-marker

I have a look, but I want to display my TRK file and t1 data, I don't know what to change?Can you help me @arokem

DIPY examples/tutorials use data that is fetched directly from a few sites that host the data of interest. If you want to adapt the examples to use your own data, then, instead of using the fetch_* and read_ * commands, you will need to read your local files.

In order to read your t1, you will need to use NiBabel as in
https://nipy.org/nibabel/nibabel_images.html

something like

import nibabel as nib

t1_img =  nib.load(your_t1_filename)
t1 = t1_img.get_data()
fig = plt.figure()
plt.imshow(t1, cmap="gray")

Similarly, to show your TRK file, you can do something similar to this

from dipy.io.streamline import load_tractogram
from fury import window, actor

tractogram = load_tractogram(your_trk_filename, your_t1_filename)
streamlines = tractogram.streamlines
renderer = window.Renderer()
stream_actor = actor.line(streamlines)
renderer.add(stream_actor)
window.record(renderer)

If you want to display your streamlines and your T1 at the same time, you will need to add both actors to the same scene, adding to the above something like

t1_actor = actor.slicer(t1)
renderer.add(t1_actor)

Jakob Wasserthal

@wasserth

Hi, is there a function in dipy to resample streamlines to make all points have a specific distance (in mm) from each other? So longer streamlines would have more points than shorter streamlines.

Wade Weber

@wadeatUT

Hi all. When I convert my Freesurfer apar+aseg back to native space to use as my WM label file in DIPY, it won't let me reslice it. I used Freesurfer's recomendations to go from Tal back to native space and to convert to NIFTI. When I load that file into DIPY and use DIPY's reslice command (from 0.8mm to 2mm), it looks like it's processing, but the shape and zooms haven't changed. Is there something special that I have to do with data converted from Freesurfer? Thanks!

zlf123-marker

@zlf123-marker

@jhlegarreta Thank you very much

Shreyas Fadnavis

@ShreyasFadnavis

Hi @wadeatUT, Can you share the piece of code where you are reslicing the data.. I don't think anything special is needed.. it would be helpful to see what the exact issue is!

Wade Weber

@wadeatUT

@ShreyasFadnavis Thanks for the help. Here's a screenshot of my Jupyter Notebook:

@ShreyasFadnavis Also, if I change "new_zooms" to (3., 3., 3.), the new vox size comes out as (1.2, 1.2, 1.2) instead of 3mm iso? Weird.

Shreyas Fadnavis

@ShreyasFadnavis

@wadeatUT I guess this line is missing: zooms = orig_t1FS_img.header.get_zooms()[:3] and should solve the issue!

Shreyas Fadnavis

@ShreyasFadnavis

Hi everyone! The DIPY Workshop 2020 Early Bird registration ends in 3 days! To all those interested, visit: https://workshop.dipy.org/ | for any questions mail us: dipy@python.org and for updates, follow us on Twitter: @dipymri !! Looking forward to see you in March at Indiana University Bloomington!

Eleftherios Garyfallidis

@Garyfallidis

:star2: :star: :rocket: :rocket: :rocket: :rocket: :rocket: :rocket: :star: :star2:

Eleftherios Garyfallidis

@Garyfallidis

Thank you for sharing @ShreyasFadnavis

Serge Koudoro

@skoudoro

:+1: :+1: :+1:

zlf123-marker

@zlf123-marker

@jhlegarreta Traceback (most recent call last):
File "C:/Users/20190401-1/Desktop/image-registration/viz_t1_dti.py", line 55, in <module>
tractogram = load_tractogram('AF_left.trk',example_filename)
File "D:\Anaconda3\lib\site-packages\dipy\io\streamline.py", line 129, in load_tractogram
if not is_header_compatible(filename, reference):
File "D:\Anaconda3\lib\site-packages\dipy\io\utils.py", line 233, in is_header_compatible
reference_2)
File "D:\Anaconda3\lib\site-packages\dipy\io\utils.py", line 195, in get_reference_info
voxel_order = ''.join(nib.aff2axcodes(affine))
TypeError: sequence item 0: expected str instance, NoneType found

@jhlegarreta It's running like an error

Francois Rheault

@frheault

@zlf123-marker is your affine all zeros?

Scott Trinkle

@scott-trinkle

quick visualization question. is there a way to get different segments of individual streamlines to have different colors? right now the default "directional" colormap seems to give the whole streamline the same color based on the overall direction, but I would like to see how the direction changes across long streamlines

zlf123-marker

@zlf123-marker

@frheault

not all zeroes

zlf123-marker

@zlf123-marker

Hello everyone, I want to know how to display TRK file and t1 file at the same time?

Everybody, I want to know how to display my TRK file and t1 file at the same time?

Eleftherios Garyfallidis

@Garyfallidis

@scott-trinkle yes you can. You can set the colors for every point of the streamline.

@zlf123-marker are your t1 and TRK in the same coordinate system? If yes then see https://dipy.org/documentation/1.0.0./examples_built/viz_advanced/#example-viz-advanced and dipy_horizon. If not then you have to do registration first.

zlf123-marker

@zlf123-marker

Ok, thank you very much

paras sethi

@dreadysharp_twitter

Hey guys.. I wanted to read about how blockwise 3d denoising has been implemented ..

But I see 2 methods nlmeans and non_local_means different

Where communities thrive

People

Repo info

Activity