# Orientation convention in dcm and nifti

**URL:** <https://discourse.itk.org/t/orientation-convention-in-dcm-and-nifti/4621>\
**Category:** Beginner Questions\
**Created:** [November 29, 2021, 9:25am UTC](https://discourse.itk.org/t/orientation-convention-in-dcm-and-nifti/4621 "2021-11-29T09:25:26Z")\
**Posts on this page:** 4\
**Page:** 1

<div class="post-metadata">

**Author:** ![BayesianMonk](https://discourse.itk.org/letter_avatar_proxy/v4/letter/b/54ee81/32.png) [@BayesianMonk](https://discourse.itk.org/u/BayesianMonk)\
**Post date:** [November 29, 2021, 9:25am UTC](https://discourse.itk.org/t/orientation-convention-in-dcm-and-nifti/4621/1 "2021-11-29T09:25:26Z")

</div>

Hi everyone,  
I am looking for the orientation convention in nifti and dicom.  
It is well known that in axial, nifti uses the RAS orientation, while dicom uses LPS orientation.  
But what about coronal and sagittal orientation?  
I think that in dicom, it is LIP for coronal, and PIL for sagittal.  
Could anyone validate this and share the results for nifti?  
Best regards

BM

---

<div class="post-metadata">

**Author:** ![dzenanz](https://discourse.itk.org/user_avatar/discourse.itk.org/dzenanz/32/1093_2.png) [@dzenanz](https://discourse.itk.org/u/dzenanz)\
**Post date:** [November 29, 2021, 9:04pm UTC](https://discourse.itk.org/t/orientation-convention-in-dcm-and-nifti/4621/2 "2021-11-29T21:04:10Z")

</div>

> [@BayesianMonk](#):
>
> nifti uses the RAS orientation

This assumes identity orientation matrix.

> [@BayesianMonk](#):
>
> I think that in dicom, it is LIP for coronal, and PIL for sagittal.

This is the first time I hear that there could be a “standardized” orientation for sagittal and coronal scans.

Anyway, you should examine the orientation matrix and that will tell you what orientation it is in. See [docs](https://github.com/InsightSoftwareConsortium/ITK/blob/047fa6934ac32711d1d7b84fa6eb81c72db01960/Modules/Core/Common/include/itkOrientationAdapterBase.h#L35-L40) and [implementation](https://github.com/InsightSoftwareConsortium/ITK/blob/8801247049712d50528e4651a0d000892b7eb22a/Modules/Core/Common/src/itkSpatialOrientationAdapter.cxx#L33).

---

<div class="post-metadata">

**Author:** ![BayesianMonk](https://discourse.itk.org/letter_avatar_proxy/v4/letter/b/54ee81/32.png) [@BayesianMonk](https://discourse.itk.org/u/BayesianMonk)\
**Post date:** [November 30, 2021, 7:58am UTC](https://discourse.itk.org/t/orientation-convention-in-dcm-and-nifti/4621/3 "2021-11-30T07:58:48Z")

</div>

Hi,  
Since there is a convention for axial scans, I was wondering that there might be one for coronal and sagittal as well.  
At least for dicom, it seems that there is such a convention since most of the cases I had to work with were oriented LIP/PIL for coronal/sagittal. Furthermore, it seems that some viewer use a convention because Mango/ITK SNAP/3D SLICER/RADIAN display the same cases with conflicting orientation: some are flipped left/right or up/down, others are 90° rotated,…  
I’m trying to get the top of this and have something really consistent, so a convention would be useful here!  
But I am happy if there is no convention, I can as well define mine for my use case 🙂

---

<div class="post-metadata">

**Author:** ![dzenanz](https://discourse.itk.org/user_avatar/discourse.itk.org/dzenanz/32/1093_2.png) [@dzenanz](https://discourse.itk.org/u/dzenanz)\
**Post date:** [November 30, 2021, 3:00pm UTC](https://discourse.itk.org/t/orientation-convention-in-dcm-and-nifti/4621/4 "2021-11-30T15:00:41Z")

</div>

> [@BayesianMonk](#):
>
> most of the cases I had to work with were oriented LIP/PIL for coronal/sagittal

This might be the consequence of working with data coming from the same institution, which uses the same manufacturer’s software and same settings.

> [@BayesianMonk](#):
>
> I can as well define mine for my use case

I think you will need to do that. I recommend mirroring Slicer.
