• mrivis version: 0.2.7
• Python version: 3.6.5
• Operating System: Fedora 27

Description

Hi Raamana, I was trying to do this with @lalet:

>>> from mrivis import checkerboard
>>> checkerboard('sub-01/anat/sub-01_T1w.nii.gz', 'sub-01/anat/sub-01_T1w.nii.gz', patch_size=10)

And this happened:

Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/glatard/.local/lib/python3.6/site-packages/mrivis/mrivis.py", line 92, in checkerboard
    **mixer_params)
  File "/home/glatard/.local/lib/python3.6/site-packages/mrivis/mrivis.py", line 398, in _compare
    display_params, **kwargs)
TypeError: _generic_mixer() takes 3 positional arguments but 6 were given

Any hint?

• deriving from Carpet, create FuncCarpet with defaults tailored to most neuroimaging applications, such as taking tissue segmentation masks, default blurring on, etc

• deriving from Carpet, create DiffusionCarpet, similar to above with useful defaults

• merge from #8 disappeared for some reason :)
• dig it or merge it back in

• 'middle' returning one middle slice from the views selected
• option to select slices within an ROI (in_roi='seg_image.nii')

• SlicePicker from 4D volumes
• e.g. to check alignment , or producing a collage

update mrivis docs to note that all my packages assume orientation is encoded correctly

find ways not to assume it.

With cross platform support in mond

Right now, it only shows alignment-related visualizations.

Add more visualizations for all features, including

• Carpet
• Collage
• SlicePicker

• to unroll 4D data over the 4th dim,
• optional input of ROI set or mask
• option to cluster the rows (spatial/voxel-wise) and/or columns (time or gradients),
• as well as select specific rows and columns
• display options include coloured lines representing ROI/voxel groups

• mrivis version: 0.2.8
• Python version: 3.4
• Operating System: CentOS7

Description

from mrivis import checkerboard
plot = checkerboard('T2w_acpc.nii.gz','T2w_acpc_centos7.nii.gz',patch_size=10)
plot.savefig("T2w_tilt.png")

These two files are output from two operating systems.   
They are completely off w.r.t alignment and we tried to visualize the differences using checkerboard. 
But no checkerboard is visible in the output image.  
Thanks for fixing the other issue, but it would be greatly appreciated if you have any hint on this one. Please find attached the output image.

@glatard

• composite layering of stats (as in visualqc)
• mixing multiple variations of carpet (temporal derivatives etc)
• exploiting multiple color channels to ingest useful info

• Add real world examples
• show the effect of different number of ROIs
• show the effect of different clustering metrics

Here I will write my review for the JOSS submission: openjournals/joss-reviews#897

General impression

The purpose of the project is simple but very important and can be divided into two parts:

1. The comparison of two 3D MRI images. This is achieved with the Alignment check functions.
2. The creation of adjustable visualization collages for 3D MRI images, as well as the creation of carpet plots for 4D images. This is achieved with Visualization Classes.

Alignment check between MRI images is very important in neuroimaging. Reducing the similarity between two images to a simple number (e.g. using a distance function) might be accurate but is often times not useful as it doesn't provide information about the source of the discrepancy. Using a visual approach, and in this case, providing multiple ones is a very nice solution to this problem!

The creation of MRI collages is also very important for the field of neuroimaging. A collage allows the investigation of a 3D image (with many 100 of slices in three directions). All at once with a quick glance. Even though this is important, there are only a few open source software that provides this functionality, but none does it as simple and flexible as mrivis! The possibility of only plotting slices with a particular non-zero density or containing a given ROI is really great! A similar "one view check" can be achieved for 4D images, using the Carpet class!

My general impression of the repo is very good! The repo has an informative README.rst file, a very detailed documentation homepage, and provides important files, such as LICENSE AUTHORS.rst, CODE_OF_CONDUCT.md, CONTRIBUTING.rst, etc.

The code is well written, uses PEP440 standard (enforced with flake8) and is well commented.

The installation of the software is very easy. mrivis can be installed very quickly via pip and all required dependencies are listed in requirements.txt.

To quickly test mrivis and getting used to its functionality, this repo provides some example files as well as a demo notebook. This is very good, as it allows a user to quickly learn how to use mrivis.

Review criteria

Concerning the review criteria under https://joss.readthedocs.io/en/latest/review_criteria.html:

• Software license is Acceptable
• Documentation is very good
• A statement of need is clearly defined
• Installation instructions are Good
• Example usage are included (with example data)
• API documentation is Good
• Community guidelines are provided
• Functionality is Good (minor bug issues mentioned below)
• Tests are Good (done with Travis-CI)

Additional Comments

I haven't reviewed the MATLAB code section in the matlab_code folder, as it contains the note: This is no longer maintained! The python version has more options and is recommended.