CGAL users discussion list

["Audette, Michel" <>]

Title: ellipsoid fitting: incorporating a known or assumed rotation of principal axes.

Dear all,

I have a student who is using CGAL software to perform an ellipsoid fitting in order to model the bite of a surgical tool whose mouth opens and closes. The fit that we get when the mouth is open appears to be a bad fit, unnecessarily large, in part because the ellipsoid seems to want to retain a standard orientation of principal axes, when in fact it would be better to give the fitting algorithm some other assumed orientation coinciding with the plane of symmetry of the open mouth, and a 3^rd axis normal to this plane. The fitting would be much more compact, and coincide with the actual bite of the tool, if we were able to tweak the orientation of principal axes.

I can send an image of our results, to anyone who responds, to make things clearer.

Is it possible to do this with the CGAL fitting software? Please advise.

Thanks for your kind consideration.

Best regards,

Michel Audette, Ph.D. 

Innovation Center Computer Assisted Surgery (ICCAS) 

Philipp-Rosenthal-Strasse 55

Leipzig, Germany

Phone: ++49 (0) 341 / 97 - 1 20 13

Fax: ++49 (0) 341 / 97 - 1 20 09