Scaling down an irregular mask


#1

Hi,

I have a mask of a irregular shaped tumor and I want to scale down this mask to a certain degree so that I can exclude the periphery and measure the core of the tumor. When I use the scale down feature of the mask there is one center point instead of multiple center points. This causes it to miss areas of the tumor that are irregular in shape.
An example is I have mask C but when I scale it down it does c missing the top portion of the actual image. What I want is C.
I hope this makes sense.

Thanks,
Darren


#2

I think what you want to scale down is the ROI (Region of Interest - your tumor) and not the whole mask image, right?

I would work with a ROI and then scale down the ROI.

From your mask you can easily create a ROI.

You can create a selection from the mask image with Edit->Selection->Create Selection.

Then you can add this selection to the ROI Manager (press t).

Select this ROI in the ROI Manager and then scale this selection (centered) down with Edit->Selection->Scale...

If you have multiple areas in the mask you get a composite selection which you can seperate in the ROI manager with: More->Split

Finally you can measure the selected ROI areas on the original image (select the image and select one or more ROI’s in the ROI manager - press action Measure)

ROI Manager docs:

https://imagej.nih.gov/ij/docs/guide/146-30.html#sub:ROI-Manager


#3

Hi @dkwchan and welcome to the forum!

I think scaling is the wrong approach here.

It seems what you need is to shrink your selection, which you can do using Edit > Selection > Enlarge… with a negative number. See also the documentation:

Enlarge…
Grows an area selection by a specified number of pixels. Enter a negative value to shrink the selection.


#4

I think shrink is where I want to go but it doesn’t shrink. I have internal holes in the tumor as places to exclude the ROI and I tried removing them so what you have is a mask that has no areas being excluded but the it isn’t shrinking.


#5

I can send you the image (which is huge) and the ROI if need be


#6

Hi imagejan! Thanks for the idea of shrinking the image but for some reason it is unable to shrink the image when I put a negative value in the enlarge box. My image is quite huge (~1GB). The ROI contains the overall mask of the tumor and excludes holes where tissues was either ripped or contain vacuoles. Thanks!