Problems with coloration on edges of blobs

I am new to using ImageJ, and my colleagues and I are trying to use it as a way of using coloration to represent a variable that ranges from 0 to 2. Briefly, we are dealing with a spatially three dimensional image for which each pixel, in addition to x,y,z values, has a value of this variable ranging from 0 to 2 (I will call this variable “A”). The data is read in via a tif file with a stack of images. I have two images below to demonstrate the problem, one a slice from the tif file and one a slice in Volume Viewer. There is a large blob and a small blob in the data.

In the data, the smaller blob contains almost exclusively pixels with A=2. Thus, as I understand it, under thermal coloring all pixels should be red (in the example of Thermal coloration system). However, the edges of the blob are green. I am not sure where this green outline is coming from. There is definitely not an A<2 outline in the data that I am using.

My speculation is that the green outline represents some sort of averaging or interpolation between the pixels with A=2 and the surrounding pixels for which A=0. For my purposes, this averaging misrepresents the data by making it appear that there is a shell of pixels with A<2. However, I haven’t been able to figure out how to stop this. An example of red with green outline is shown below:

Slice in Volume Viewer, Thermal Color option

This problem also appears looking at a slice from the tif file, with the spectrum overlay for coloration (again, a shell with distinct properties shows around the edges of the small blob):

Thank you very much in advance for any suggestions.

What kind of image or data (datatype in general) do we see? 8-bit, 16-bit, 32-bit? - The image information could help.

Did you control the values of the pixels where this wrong display occurs. Are they still correct?

Thank you for your reply, I apologize, I didn’t see it until today. I have found a work-around for the problem that avoids interpolation issues, so I am going to delete this post. Thank you very much for your time in responding.

1 Like