Fiji: Isotropic issue after reslicing

Hi everyone

I have a stack image (845 slices), with 2048X2048 pixels for each slice.
It is a mouse brain image and the raw image was taken at horizontal plane.
I wanted to reslice horizontal plane to Coronal plane so I used the “reslice” function. (Make XYZ to XZY).
However the images after reslicing looks weird and not isotropic (not uniform).


Here you can see some areas of the image (after reslicing) are clear but some areas are not.
Does anyone know what may cause this? Or any ideas about how to do a better reslicing? Thanks
(The resolution became 2048*845 pixels due to the limitation of the original Z-stack number, but I think this shouldn’t be the reason that the image is not uniform?)

It might help to know a little bit more about how the image was taken, as the two well focused areas appear to be symmetrical, suggesting some sort of radial symmetry perhaps? I am not sure.

1 Like

Hi, thanks for the feedback!
The images were taken from the superior side to inferior side by a light sheet microscope.
Like the example here:
Picture3

Hence, the original images are a stack of slices in horizontal plane.
Here is one of the slice from original images:

As you can see, the image here is very clear and doesn’t have the unfocused issue.
I am not sure why some areas became unfocused after reslicing the images to a different orthogonal plane (in this case from horizontal plane to coronal plane)
planes (1)

And yes, you made a point! The focused areas are symmetrical suggesting a radial symmetry issue, but I have no idea what may cause this. Is it the algorithm that caused this, or is it a physical limitation of the image itself? I think I will try orthogonal transformation at another programming language (maybe Python) to see if it is the same. Any suggestions? Thanks

1 Like

Hey @lyonleu ,

I assume you used this dialog when reslicing in Fiji, correct?
image

Can you try changing the output slicing and the “avoid interpolation” checkbox?

Furthermore, how is the image calibrated before and after reslicing? You can see this in Image > Properties:
image

Let us know if this helps!

Cheers,
Robert

3 Likes

Hi Robert,
Thanks a lot for the feedback!
Yes, I used the exact “Reslice” dialog like the one you showed.
I did try to change the output spacing and the “avoid interpolation” option, but they all have the partially unfocused issue.
Here is the properties before and after reslice.
Before:
Picture6
After:
Picture5

Looks like it doesn’t change that much? Thanks

What is the pixel/voxel size of your data? Can you enter it in the Image > Properties dialog before reslicing?

2 Likes

Hi Robert,
Thanks for the suggestion.
So I entered the pixel/voxel size to the properties before reslice.
The properties before reslice:
Picture10
The properties after reslice (Output spacing set as 9 with interpolation since I ran out of memory):
Picture11
The resliced image has a better aspect ratio now. Thanks for this!
However it still has unfocused issue with the similar patterns like the previous one…

Is there any chance there was processing done prior to your reslicing? Any kind of deconvolution, alignment from images being taken from multiple angles, or anything like that?

Kind of at a loss though, have not dealt with any lightsheet data directly.

1 Like

Comparing your earlier screenshot:

With the more recent

I would say the isotropy issue is gone.

What kind of imaging technique was used to acquire the data? “blurry” deep inside tissue is quite common, e.g. in light sheet microscopy…

1 Like

Hi MicroscopyRA,

Thanks for the comments.
I didn’t do any processing before slicing.
The only thing I did is “import image sequence” and made it to a stack.
I am not sure if I accidentally involved some processing here?
The image was taken by light sheet microscope. I am also not an expert regarding how it works, but maybe I can consult some experts (perhaps some alignments were done while the image was taken?). Thanks

Hi Robert,

Thanks for the comments.
The image was taken by light sheet microscope.
I think the more recent one looks a bit better, but there are still unfocused areas.
The tissue was cleared so basically it was close to “transparent” while being imaged.
The original images look really good and are all on focus. So I am really confused about the reslicing results.

I actually tried to reslice the image by Python earlier, and sadly the result is the same with the similar pattern. I started to think maybe it is a physical issue while taking the images…Thanks

1 Like

If you add the ImageScience plugin repository (I think that’s the name, not in front of a computer at the moment), then the “TransformJ” plugin should do the trick. Tell it to rotate either X or Y by 90 degrees, with isotopic resampling.

https://imagej.net/TransformJ

1 Like

Hi Marc,

Thanks for the suggestions!
I tried the “TransformJ” plugin, and chose “TransformJ Rotate” with “Resemble isotopically”.
It is a straightforward and convenient way to reslice stacks! However the result is still the same, with some unfocused areas at the middle and the edges.

This seems to be really tricky…Thanks