Wrong volume measurement

Hello,

I have used the volume fraction plugin of BoneJ to estimate the volume fraction in the stack of the micro-CT image of an open-cell foam. In the stack, there are 100 slices, and the pixel size is 2.25597 um. When I run the volume fraction plugin in BoneJ, although the value of volume ratio seems reasonable, the value of the total volume is WAY lower than what I expect. The dimension of the images is 1800x1200 pixels, and I have 100 slices in the stack, so I expect the total volume to be 1800x1200x100x2.25597x2.25597x2.25597 um3 = 2 480 011 595,39 um3. However, the boneJ result for the total volume is 97638.38 um3. Do you think I am missing something or the boneJ output is wrong ?

Thanks,
Reza

Could you zip and post the image so we can check?

You should make sure that the pixel spacing is set for all 3 dimensions by looking at Image > Properties.

Hello,

I checked the properties as you mentioned, I figured that the voxel depth was not the same as pixel width and pixel height. Do you think if it could influence the Tb.Th measurements ?

Thanks,
Reza

Thickness ignores the pixel spacing during sphere fitting, because it assumes pixels are on an isotropic grid. It should have warned you if your pixel spacing was anisotropic (if it didn’t, that is a bug). Tb.Th calibrates back to real units from pixel units using only one of the dimensions - I think (without checking the code) it’s the x pixel spacing. In general avoid running Thickness on anisotropic pixels.

Thanks. I imagine it did give me a warning. I re-ran Tb.Th and volume fraction measurements on some of my dataset (this time with voxels having identical values in all 3 dimensions), and the final result for Tb.Th and the volume ratio turned out the same. As expected, the total volume and the bone volume values changed.