# Py RADIOMICS query

**URL:** <https://discourse.slicer.org/t/py-radiomics-query/5443>\
**Category:** Support\
**Tags:** pyradiomics\
**Created:** [January 21, 2019, 2:41pm UTC](https://discourse.slicer.org/t/py-radiomics-query/5443 "2019-01-21T14:41:18Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![RadioQuest](https://avatars.discourse-cdn.com/v4/letter/r/85f322/32.png) [@RadioQuest](https://discourse.slicer.org/u/RadioQuest)\
**Post date:** [January 21, 2019, 2:41pm UTC](https://discourse.slicer.org/t/py-radiomics-query/5443/1 "2019-01-21T14:41:18Z")

</div>

Hi…  
I am working on TA to study changes in a longitudinal study.  
I am doing 2 step registration process - 1. affine from A to B and later Bspline using this affine as a linear transform.  
However, I have following queries

1. Is it necessary to select histogram matching for TA (I have already performed N4 bias correction)
2. If I have to select this option, shall I choose this during affine transformation or non-rigid part of registration?  
Please note brain scans B/C etc are distorted post surgery.
3. And does baseline scan (A) need any registration ? as it will be used as a template and also control?  
Thanks in advance.  
Kind Regards,  
RQ

---

<div class="post-metadata">

**Author:** ![pieper](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/pieper/32/8_2.png) [@pieper](https://discourse.slicer.org/u/pieper)\
**Post date:** [February 6, 2019, 6:13pm UTC](https://discourse.slicer.org/t/py-radiomics-query/5443/2 "2019-02-06T18:13:12Z")

</div>

Hi -

Probably best to post to [the pyradiomics list](https://groups.google.com/forum/#!forum/pyradiomics) if you haven’t already.

My first guess at the answers are:

1. still a good idea
2. probably use the untransformed data to make the histogram (or mask it if the post-surgery images have artifacts that would skew the histogram)
3. No, not if it’s the reference space.