"Show Info" Depth

Hi all

I was curious about something in the Image->“Show Info” window. The “depth” seems to correspond to “number of slices” * voxel depth.

However in microscopy voxel depth usually corresponds to a “stage step size”, so should depth correspond to “number of slices-1”*voxel depth?? IE if you have 2 slices, and a step size of 1 micron, you’ve traveled 1 micron. Or am I thinking of this wrong?? Are there applications where you’d want a true “voxel depth” instead of a “step size”??

Brian

I would say that:

image depth = number of slices * voxel depth

in the very same way as:

image width = number of pixels along x axis * pixel width
image height = number of pixels along y axis * pixel height

It mainly depends on how you define the border of your sampled area: does the border coincide with the outermost sample points of your image, or is the border spaced in between grid positions, i.e. 0.5 pixels off from the border pixel positions. So it all comes back to the old discussion of how to define ROIs and boundaries, no?

1 Like

Good day Jan,

sorry to mention it again and again and I’m sure you are aware of it:

There is no pixel width, height or voxel depth, there are only the corresponding spacings.

A discrete signal consists of numbers and they have no (temporal or spatial) extension, they are mathematical points. The misunderstanding that pixels (picture elements) are areas comes from the problem of displaying point-like values. The cheapest way of doing this is by zero order interpolation which leads to areas of constant color. Of course this interpolation is incorrect – but easy to accomplish.

Best regards

Herbie

Yes, @anon96376101, thanks for insisting, that’s correct, and from a signal processing point of view, I fully agree.

The problem is that in the ImageJ 1.x user interface, there are these terms, literally Pixel width, Pixel height, and Voxel depth:

The whole legacy interface of ImageJ essentially builds on the idea of pixels having a defined area (or volume), hence all the confusions from the user perspective. See also the endless discussions about ROI areas and outlines, and several bugs with misplaced point ROIs and some such in the past.

1 Like

In reality the dialogue should say “vertical inter-pixel distance”, etc.
Brian is correct about the calculation of the stage movement, as a single image has thickness of 0. And I also agree it comes to the same problem of the ROI positioning. Will people get confused? Of course! :smile:

1 Like

So I guess the question is what can be done??

I don’t think anything should be changed on the core interface. It has been that way for years and not something I would want to tinker with.

However I wonder if it possible to add to and/or modify the text in the “Show Info” print out.

Perhaps I could look into adding a second field called “stage movement”?? Perhaps it could say

Depth (sum of voxel depth) = (num pixels * voxel depth)
Stage Movement = ( (num pixels-1) * voxel depth)

I guess this gets complicated because there will be occasional situations of uneven stage movement, and in that case Stage Movement should be (position n - position 0).

Let me know what people think. Thanks.

1 Like

"Will people get confused? Of course! "

“So I guess the question is what can be done??”

There are more inconsistencies of this kind with ImageJ but I’m convinced we can live with them, provided we are willing to recognize them as such.

Back to work …

Herbie

1 Like