Let me first say that I am perfectly capable of using other methods to do what cannot be done in the Caliper Ellipse Fit. The point I am stating in this post is that without rotation in the results, the function is pretty much a dead end, as there is little you can do with the one point it returns.
IMO the Caliper Ellipse Fit needs to return at least three additional results. 1. Rotation angle of the ellipse (wide or narrow axis); 2. Two points on the wide axis perimeter; 3. Two points on the narrow axis perimeter.
The idea behind the Caliper funtions is that they return points which can be fed forward into other calipers to do all sorts of things. I think the Ellipse Fit abandons this idea in its current form.
