BoneJ Particle Analyzer NaN in Ellipsoid Major, Int. and Minor radius

[Kalle Lehto]

Hi,
what could be the reason for Not a Number (NaN) in Major, Int. and Minor radius outputs after the ellipsoid fit in particle analyzer? If I crop those NaN particles from the image stack and analyze them again with Particle Analyzer I`ll get resonalble results in ellipsoid fit without no NaN in the output.

[Michael Doube]

Hi Kalle,

Most likely it's infinitely small ellipsoids or particles for which ellipsoid fitting failed. Is there a common shape among the particles with NaN radii?

Michael

[Kalle Lehto]

Hi Michael,
I think that there is no any common shape considering those NaN particles. Volume and area for these NaN particles is also large enough. In the preprocessing I already excluded all particles less than 90 unit^3, thus, segmentation errors such as noise and other unwanted artefacts are already exluded. It is also weird that I can analaze these NaN particles if I exclude other particles from the analysis.

Kalle

[Michael Doube]

Hi Kalle,

The code has a clue. The default value for the result entry is NaN. If the ellipsoid fitting for a particle returns null, then NaN is entered.

Check whether an ellipsoid is generated for each particle by turning on the 3D output 'show ellipsoids' and one of the other values, e.g. 'show axes'. If there are missing ellipsoids for some axes then we have a culprit: ellipsoid fitting failed. Could you please check that on your data.

I can see that there is another possibility in the ellipsoid fitting class FitEllipsoid, which is that the radii are calculated from eigenvalues, and maybe that maths is broken. If I remember correctly the Petrov fit that BoneJ uses does not constrain the result to an ellipsoidal quadric and other quadric surfaces may result from the fitting, if the input data are not particularly ellipsoidal in shape. That would be a strong hint that modelling such particles as ellipsoids is inappropriate. See this for a list of other potential quadric surfaces: https://en.wikipedia.org/wiki/Quadric

Best regards,

Michael

--
You received this message because you are subscribed to the Google Groups "BoneJ Users and Developers" group.
To unsubscribe from this group and stop receiving emails from it, send an email to bonej-users-and-dev...@googlegroups.com.
To post to this group, send email to bonej-users-a...@googlegroups.com.
Visit this group at http://groups.google.com/group/bonej-users-and-developers.
For more options, visit https://groups.google.com/d/optout.