Compute Cross Sectional Area

[william-silversmith]

Radius is good, but cross sectional area is even better. We should find a way to do something like a windowed moving average vector as you walk along the skeleton and compute the cross sectional area of the orthogonal plane. We'll have to experiment to see what the best vector tweening algorithm is.

[william-silversmith]

In discussions with FlyWire users, they indicated that one interpretation that is useful for simulation is the "equivalent" radius as many models use e.g. cylinders to approximate the wire. That is, A = pi * r^2. However, I think since this is trivially computable from the cross sectional area, it makes little difference how it is presented.

[william-silversmith]

One approach to doing this is "sampling" where we tween along the basis vectors sampling every λ/4 or more frequently. This makes it possibly more expensive but mathematically easier to evaluate the contribution of each intersected voxel. However, a more precise calculation will likely be able to achieve similar precision with less repetitive work.