Biomedical Engineering Reference
In-Depth Information
the agglomerative clustering. This clustering method is described more precisely in
the following. Finally an idealized hip cup prosthesis is designed using the resulting
universal geometry of the edge of the acetabulum.
4.2 Preliminary Investigation
4.2.1 Method
The basis for the use of the deduction of the universal geometry is the superposition
of the trimmed 3D models. Therefore a suitable method has to be found. Thus a
preliminary investigation was executed [
18
]: In this study two different superposi-
tion methods, the so called Best Fit and the Reference Point System (RPS), were
compared. Both of them are provided by the commercial software Geomagic Qual-
ify 2012 (Geomagic, USA). The analyses are realized with 11 segmented pelvis
geometries prepared as described before by the design chain.
The generated data are compared in the same way. First one geometry is selected
randomly from the 11 segmented pelvis geometries used for both superpositionmeth-
ods. This geometry is defined as reference for the following analysis. The remaining
10 geometries are aligned as test with the reference by using both methods. In this
context a geometry defined as test is allowed to move against the geometry defined
as reference which is fixed. After each superposition the
s
tandard
d
eviation of the
g
eometrical deviation (SDG) is determined. By means of the SDG the superposi-
tion quality can be assessed, because this value implies the information about the
deviation over all the geometries.
For the RPS method 21 points are located by hand on reproducible points of the
acetabular area. The RPS method is performed seven times with a varying number
of utilized points (6-21 points) in order to examine the influence of the count of
points on the superposition quality. For each group of points the average of the SDG
is calculated. The Best Fit is executed once. In this case as well, the average of the
SDG is calculated for all superpositions.
4.2.2 Results
In Fig.
5
the mean SDG after using the RPS is plotted against the varying number of
points. Furthermore, the mean standard variance after using the Best Fit is illustrated
as constant dashed line for a better comparison with the results of after the RPS. It
can be noted that an increasing number of points leads to a decrease of the mean
standard variance and consequently to a higher quality of superposition. The value
of the Best Fit, however, is not reached with RPS.
Thus, the Best Fit seems to be the more suitable method for the superposition of
the pelvis geometries and is selected for the derivation of the universal geometry.
