Multi-Shell Free Water Elimination Using MT-CSD

[rutgerfick]

• [x] includes function to remove specific isotropic signal contributions from fitted MT-CSD signal reconstruction
• [x] test for the correct volume fraction removal.

• [ ] Examples to apply this to multi shell data (MAP-MRI)

  SS3T saved for later since it requires comparison with SS2T (WM+CSF) to check if it actually estimates CSF fractions better.

[coveralls]

Pull Request Test Coverage Report for Build 425

• 30 of 32 (93.75%) changed or added relevant lines in 2 files are covered.
• 4 unchanged lines in 2 files lost coverage.
• Overall coverage increased (+0.3%) to 83.323%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
dmipy/core/fitted_modeling_framework.py	12	14	85.71%
<!--	Total:	30	32	93.75%	-->

Files with Coverage Reduction	New Missed Lines	%
dmipy/utils/tests/test_spherical_convolution.py	1	95.35%
dmipy/distributions/tests/test_bingham.py	3	86.3%
<!--	Total:	4		-->

Totals
Change from base Build 416:	0.3%
Covered Lines:	4516
Relevant Lines:	5285

---

💛 - Coveralls

[codecov-io]

Codecov Report

Merging #26 into master will increase coverage by 0.01%. The diff coverage is 90.62%.

@@            Coverage Diff             @@
##           master      #26      +/-   ##
==========================================
+ Coverage   82.35%   82.36%   +0.01%     
==========================================
  Files          61       62       +1     
  Lines        5253     5285      +32     
  Branches      614      617       +3     
==========================================
+ Hits         4326     4353      +27     
- Misses        763      769       +6     
+ Partials      164      163       -1

Impacted Files	Coverage Δ
dmipy/core/tests/test_return_filtered_signal.py	100% <100%> (ø)
dmipy/core/fitted_modeling_framework.py	65.89% <78.57%> (+1.13%)	:arrow_up:
dmipy/utils/tests/test_spherical_convolution.py	95.12% <0%> (-4.88%)	:arrow_down:
dmipy/distributions/tests/test_bingham.py	91.04% <0%> (-2.99%)	:arrow_down:
dmipy/tissue_response/white_matter_response.py	72.15% <0%> (ø)	:arrow_up:

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update e1ba0bd...93d8b90. Read the comment docs.

[rutgerfick]

The kernels need to be re-estimated from a group of noisy response functions for every SNR. This is likely the reason why noisy simulations didn't give the same results when estimating RTOP etc. after removing the noiseless csf kernel.