Result and Discussion - Segmentation of Hand Bone for Bone Age Assessment - page 110

analysis procedures are required, and this definitely increases the computational
efficiency if the number of user-specified parameters increases. Similarly, the
number of training samples, N TI exhibits linear incremental growth in total num-
ber of user-specified parameters while the proposed method are not affected by
training samples.
The influence of the number of targeted radiographs and training samples to
the total number of user-specified parameters has been analyzed. Now the influ-
ence of number of shape points and the number of shape point locations to the
total number of user-specified parameters are analyzed. Obviously, both number
of shape points and number of their locations are of the same amounts, therefore,
( N SP ) = ( N SPL ) , ( N SPL ) contains 2 parameters, which are the 2D coordinations
of the shape point location, and thus Eq. ( 1.2 ) is redefined to Eq. ( 1.4 ) . To examine
the influence, the other factors such as N R and N TI are assumed to be the con-
stants, C , as shown in Eq. ( 1.5 ) . Figure 4.4 illustrates that the N TNP exhibits lin-
ear-like growth as N SP increases on 100 targeted radiograph using 1,559 training
samples. This indicates that if the number of shape points is increased in order to
capture the complicated shape of hand bones when the patients ages increases, the
number of parameters that requires expert to insert increases linearly.

N TI

(4.4)

N TNP = 4 ( N SP ( I )) + 3 ( N R ) + 10

i = 1

(4.5)

N TNP = C ( N SP ( I )) + C

The increment of the parameters such as the number of membership functions,
the shape of the membership functions, and the location of the membership func-
tions also exhibit the similar linear growth effect to the N TNP similar to Fig. 4.4 .
This indicates only that, for example, if the number of membership function
increases, the N TNP grows with same pace as in N SP . However, the gist of the

7 x 10 5

6

5

4

3

2

1

0

0 10 20 30 40 50 60 70 80 90 100

1 63

Fig. 4.4 The change in N TNP as N SP increases