Page 175 - Advances in Biomechanics and Tissue Regeneration

8.5 PATIENT-SPECIFIC CARDIAC PODI COMPUTATION 171

FIG. 8.34 Registration of perturbed BV-1 mesh before registration (left side) and after registration (right side).

GA GA
5 GB 50 GB
GC GC
4 40
Mean distance 3 Minimum angle 30

2
20
1 10

(A) 0 (B) 0

0.85
GA
GB
0.8 GC
0.75
Mean shape 0.65
0.7

0.6

0.55
(C) 0.5
FIG. 8.35 Mesh quality before and after perturbation, and after registration. (A) Mean distance; (B) minimum angle; and (C) mean shape.

template, the projection of the solutions to and from the template is the most expensive operation, as 90% of the cal-
culation time is spent during that process.
In terms of solution accuracy, the ε BV R -norm obtained for the heart grid template is about 0.22, which represents a
‘ 2
39% improvement as compared to the 990-node cube template. This increase in solution quality and smoothness can
also be confirmed when a BV slice is visually compared, first against the corresponding full-scale EFG solution, as
given in Fig. 8.36, and second against the cube template solutions, as shown in Fig. 8.32. Also, these results are found
2
to benefit the pressure-volume relationship curves of the LV and RV, as their errors are 3.1 10 2 and 4.2 10 ,

I. BIOMECHANICS