Modify `pixdim` to `original_pixdim` in meta dict

[KumoLiu]

Is your feature request related to a problem? Please describe.
Follow up of #6832, after Spacing, the original pixdim is not modified in meta dict, which may cause some confusion.

Describe the solution you'd like
Would be better to update it to original_pixdim.
Similar to affine.

MONAI/monai/data/image_reader.py

Lines 279 to 281 in 8e99af5

	header[MetaKeys.ORIGINAL_AFFINE] = self._get_affine(i, self.affine_lps_to_ras)
	header[MetaKeys.SPACE] = SpaceKeys.RAS if self.affine_lps_to_ras else SpaceKeys.LPS
	header[MetaKeys.AFFINE] = header[MetaKeys.ORIGINAL_AFFINE].copy()

[IamTingTing]

Hi @KumoLiu , I'm working on this issue with @slicepaste and @einsyang723. We would like to clarify the meaning of original_pixdim mentioned in the request. Does it refer to the pixel dimension of the original image or the previous image? Like, we have an image that becomes a second image after spacing, and then a third image after spacing again. In this case, does original_pixdim refer to the dimension of the very first image or the second image?

[KumoLiu]

Hi @KumoLiu , we're working on this issue. We would like to clarify the meaning of original_pixdim mentioned in the request. Does it refer to the pixel dimension of the original image or the previous image? Like, we have an image that becomes a second image after spacing, and then a third image after spacing again. In this case, does original_pixdim refer to the dimension of the very first image or the second image?

Hi @IamTingTing, this refers to the "original image" before any operations are performed.

From your example: "We have an image that becomes a second image after spacing, and then a third image after spacing again," it specifically means the pixdim of the first image, before any spacing adjustments.

And BTW to be simple, we can only update the pixdim in the meta dict to the original_pixdim since data.pixdim is logging the correct(current) pixdim.

[slicepaste]

Hi @IamTingTing, this refers to the "original image" before any operations are performed.

From your example: "We have an image that becomes a second image after spacing, and then a third image after spacing again," it specifically means the pixdim of the first image, before any spacing adjustments.

And BTW to be simple, we can only update the pixdim in the meta dict to the original_pixdim since data.pixdim is logging the correct(current) pixdim.

Hi @KumoLiu ,
I appreciate the feedback regarding the change of pixdim to original_pixdim. However, I think maybe it would be better to ensure consistency between pixdim and affine in the code.

Currently, affine is handled as follows:

data.affine
data['affine']
data['original_affine']

Perhaps pixdim should follow the same structure, like:

data.pixdim
data['pixdim']
data['original_pixdim']

It may be more intuitive as both parameters follow the same pattern, instead of updating pixdim to original_pixdim.
What do you think about this suggestion?

[KumoLiu]

Hi @slicepaste, yes we can follow the same structure.