Preconditioning

Here I applied preconditioning on the helix using the following fitting goals Fomel (1997):

$$\begin{eqnarray} \ {substitute: } \bf m={\bf A^{-1}}{\bf p}\\ \ {\bf W} \frac{d... ...pprox& {\bf 0} \nonumber\\ \ \epsilon {\bf p} &\approx& {\bf 0} \end{eqnarray}$$ (5) (6)

In this case, a 2D PEF was estimated on the dense tracks and applied as regularization to the sparse tracks. This greatly improved the speed of convergence to create Figure 15.

 

madfig7
Figure 15 Result of fitting goal 6 applied to both sparse and dense data together usinga 2D PEF estimated on dense data.