# Extension manager for the sake of deep learning

**URL:** <https://discourse.slicer.org/t/extension-manager-for-the-sake-of-deep-learning/12015>\
**Category:** Support\
**Tags:** segmentation, dicom, deep-learning\
**Created:** [June 14, 2020, 9:57am UTC](https://discourse.slicer.org/t/extension-manager-for-the-sake-of-deep-learning/12015 "2020-06-14T09:57:30Z")\
**Posts on this page:** 2\
**Page:** 1

<div class="post-metadata">

**Author:** ![khikho](https://avatars.discourse-cdn.com/v4/letter/k/f475e1/32.png) [@khikho](https://discourse.slicer.org/u/khikho)\
**Post date:** [June 14, 2020, 9:57am UTC](https://discourse.slicer.org/t/extension-manager-for-the-sake-of-deep-learning/12015/1 "2020-06-14T09:57:30Z")

</div>

hi  
Is there Extension manager for the sake of deep learning?

---

<div class="post-metadata">

**Author:** ![lassoan](https://sea2.discourse-cdn.com/flex002/user_avatar/discourse.slicer.org/lassoan/32/13_2.png) [@lassoan](https://discourse.slicer.org/u/lassoan)\
**Post date:** [June 14, 2020, 2:58pm UTC](https://discourse.slicer.org/t/extension-manager-for-the-sake-of-deep-learning/12015/2 "2020-06-14T14:58:55Z")

</div>

3D Slicer (and its extensions) is a great tool for creating, evaluating, and deploying machine learning models for medical imaging applications, and can be used with any deep learning toolkits (Keras, PyTorch, TensorFlow, NVidia Clara, etc.).

You can use [Segment Editor](https://slicer.readthedocs.io/en/latest/user_guide/modules/segmenteditor.html) and various segmentation extensions (such as [SegmentEditorExtraEffects](https://github.com/lassoan/SlicerSegmentEditorExtraEffects), [SurfaceWrap](https://github.com/sebastianandress/Slicer-SurfaceWrapSolidify), …).for creating training data set. For segmenting hundreds of data sets, you may find [CaseIterator extension](https://github.com/JoostJM/SlicerCaseIterator) (for all kinds of 3D volumes), [AIGT modules](https://github.com/SlicerIGT/aigt) (mainly for ultrasound images) useful.

For training and quick review of results, you can use Jupyter notebooks in Slicer, using [SlicerJupyter extension](https://github.com/Slicer/SlicerJupyter). Slicer allows both qualitative visual assessment using its sophisticated 3D visualization features and computation of quantitative metrics.

For deployment - to make your deep learning models easily accessible to clinical end users - you can use [NVidia AIAA extension](https://github.com/NVIDIA/ai-assisted-annotation-client/tree/master/slicer-plugin), which makes NVidia Clara models directly in Slicer’s Segment Editor. You may also learn from some previous efforts, such [DeepInfer](https://www.slicer.org/wiki/Documentation/Nightly/Modules/DeepInfer) and [TOMAAT](https://github.com/faustomilletari/TOMAAT-Slicer).