Header has wrong orientation

[simonqueric]

Some images have wrong orientation information in the header.

See bids-rsna-lscd/sub-289846404/anat/sub-289846404_acq-ax_rec1791777852_T2w.nii.gz :

For instance, the AP axis and the SI axis are switched.

If you don't have access to the bidsified dataset, you can see dicom files in 289846404/1791777852

We have to :

• Check if the problem is limited to few subjects
• Check if the problem comes from the bidsification or the original dicom.

[simonqueric]

It also concerns axial acquisition 3060312027 of subject 1723430291

See below (AP axis and SI axis are switched) :

[jcohenadad]

Good catch!

Check if the problem comes from the #14 or the original dicom.

I would start with this check (open the image in Horos)

[Nilser3]

Thank you @simonqueric
Viewing subject 289846404 folder 1791777852 (axial):
I see that when converting the folder 1791777852 to nifti, we have 6 images and that the image in red is the one that is incorrectly encoded.

possibly because the green volume is an image acquired in axial, but as it is already close to being coronal, the encoding was erroneous (seeing the image in Horos, I also observe this problem).

For the subject 1723430291 folder 3060312027 (axial), the same event is happening:

However, looking the annotations, I note that the annotations were made using the volumes in green only, so I think we can discard the red volumes.