NUL Fabric tensor values in Anisotropy BoneJ

Good day,

I read with a lot of interest the following thread from which I could safe the information that the tensors vector from the smallest tenseur value could give the general preferential orientation.

Then I realized that the Fabric tensor values that I get in the result tab from Anisotropic were either nul or equal to one bu I still had a degree of Anisotropy (0,392)

Fabric tensor vectors

|| 0.629| 0.772|-0.094||
||-0.167| 0.016|-0.986||
|| 0.759|-0.636|-0.139||

Fabric tensor values

||0.000|0.000|0.000||
||0.000|0.000|0.000||
||0.000|0.000|0.000||
1	min5voxels	0.392	1.644	0.629	0.772	-0.094	-0.167	0.016	-0.986	0.759	-0.636	-0.139	1.302E-4	1.568E-4	2.140E-4

It is also true that the ellipsoïde is almost a sphere

Could you tell me what is going on with the Fabric tensor values ?

I study x-ray µCT tomography soil images.

Hi @sarahdata1

Perhaps @mdoube or @rimadoma would be able to assist you in this?

eta :slight_smile:

It’s not - if spherical it would look like a circle in this view, but I see an ellipse with its long axis at ~10 o’clock - 4 o’clock

DA is calculated as 1 - (smallest / largest eigenvalue).

The eigenvalues are reported here:

So in this example DA = 1 - 1.302E-4 / 2.140E-4 = 0.392

(the code that does this calculation is here)

The “fabric tensor values” reported in the log window are not particularly helpful when they get small, because it’s truncated at 3 decimal places. So you can ignore them in this case and have them recorded in the Results window instead.

I’m looking at your results and see no 1 or NULL values - could you give an example where this happens please because it may be a bug.

2 Likes

Hello Michael,

Sorry, I only looked at the log window, that’s why I tought there was only nul values (but there isn’t if you look attitude the results window…)… I figured it out before but I could not delete the post…

Thank you anyway for your answer, your expertise regarding the shape of the ellipsoïde was helpful!

Sorry again!