# Any way to parallelize radiomics feature extraction on python?

**URL:** <https://discourse.slicer.org/t/any-way-to-parallelize-radiomics-feature-extraction-on-python/12968>\
**Category:** Support\
**Created:** [August 12, 2020, 3:54pm UTC](https://discourse.slicer.org/t/any-way-to-parallelize-radiomics-feature-extraction-on-python/12968 "2020-08-12T15:54:20Z")\
**Posts on this page:** 5\
**Page:** 1

<div class="post-metadata">

**Author:** ![xlucox](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/xlucox/32/6381_2.png) [@xlucox](https://discourse.slicer.org/u/xlucox)\
**Post date:** [August 12, 2020, 3:54pm UTC](https://discourse.slicer.org/t/any-way-to-parallelize-radiomics-feature-extraction-on-python/12968/1 "2020-08-12T15:54:20Z")

</div>

Hi, I’m dealing with the extraction of a large set of radiomics feature in a large set of data, It took to much running time doing it with a simple loop in Python. I wonder if there is any way to fast it up using GPU.

On the other hand is there some library to extract radiomics features in C.

Thank you !!!

---

<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:** [August 12, 2020, 4:27pm UTC](https://discourse.slicer.org/t/any-way-to-parallelize-radiomics-feature-extraction-on-python/12968/2 "2020-08-12T16:27:37Z")

</div>

Thanks for the inquiry - we have been exploring some options like that but nothing there’s concrete at the moment. But it’s good to know there’s community interest.

For many radiomics use cases you can leverage subject-level parallelism because there are a large number of similar cases to analyze and you can run them on independent cores or compute nodes in a cluster.

---

<div class="post-metadata">

**Author:** ![JoostJM](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/joostjm/32/1091_2.png) [@JoostJM](https://discourse.slicer.org/u/JoostJM)\
**Post date:** [August 12, 2020, 4:32pm UTC](https://discourse.slicer.org/t/any-way-to-parallelize-radiomics-feature-extraction-on-python/12968/3 "2020-08-12T16:32:32Z")

</div>

In addition to Steve’s comment, subject-level parallelization is already available in PyRadiomics, via the commandline interface. To enable this, specify the `--jobs` argument with an integer value specifying how many cores to use.

---

<div class="post-metadata">

**Author:** ![thewtex](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/thewtex/32/32_2.png) [@thewtex](https://discourse.slicer.org/u/thewtex)\
**Post date:** [August 13, 2020, 5:21pm UTC](https://discourse.slicer.org/t/any-way-to-parallelize-radiomics-feature-extraction-on-python/12968/4 "2020-08-13T17:21:03Z")

</div>

The [itk-texturefeatures](https://github.com/InsightSoftwareConsortium/ITKTextureFeatures) Python package provides parallelized, fast, computation of a few radiomics features in C++.

---

<div class="post-metadata">

**Author:** ![xlucox](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/xlucox/32/6381_2.png) [@xlucox](https://discourse.slicer.org/u/xlucox)\
**Post date:** [August 15, 2020, 11:15pm UTC](https://discourse.slicer.org/t/any-way-to-parallelize-radiomics-feature-extraction-on-python/12968/5 "2020-08-15T23:15:23Z")

</div>

Thank you!!!

It was very helpful.