Full text: Proceedings of the Symposium "From Analytical to Digital" (Part 2)

In order to investigate each term by itself, we notice that (11) is 
satisfied when y is chosen in such a way that f = aD' à. The k-th 
component becomes 
fsaAg ; Km 1,2,...,M (16) 
from which we obtain 
log. (f, /a$,) 
k^ ^y M 
UK) = —————— = - log (Ë, /ad, ) (17) 
100 Ac 2kri k k 
which should be compared to (15). Here y has been given as a function 
of k, as we generally obtain a different value of y for every k. The 
complex logarithms in these expressions are easily expressed in the 
more familiar real form: 
  
Imag (f£,/8,) 
M M. abre Eee 
Real p(k) = k Luce oT arctg Sei VNIN (18) 
Imag u(k) - —- Au log T [Rea (f, /8,) * ing 0,8 3] 7 (19) 
Ls:0, 1,:2,. |. 5 K-1 
M is the period corresponding to a given spatial frequency k. L is 
fne number of complete periods involved in the displacement. Only one 
of the k possible integers L is correct. In the case when the image f 
is a pure cyclic translation of the image g, we obtain the same value 
of u for all values of k: 
| = M N= 
Real p(k) = k Lot kr Pk Oo € 94 « 27 (20) 
SUUM. 
Imag p(k) = T log a (21) 
where vw = 2kwu/M is the arctg-function in (18). Several procedures 
for finding L can be conceived. Here we choose a method which can also 
be used in the case where there are discrepancies between the images: 
For each k, y is determined for all possible values of L. A vector of 
M components (the image size) 1s constructed where the components 
indexed [u] and [p+1] (k different cases) are set equal to 1 and the 
rest are set equal to O. All the vectors are added and y chosen as the 
index of the largest component. This method is used in the following 
examples, which are typical for a large number of investigated cases 
with translations ranging from zero to half the image size. 
Example: Cyclically displaced Sine Curve: 
E(x) = sin, 21x/38; x = 1,2, ...8 
G(x) = sin 27(x121)/38; x = 1,2, canl3.®= M=2] (22) 
g(x) = sin.27(x-43)/38; X =.44 45 ....M 
M = 64 
This choice of f and g corresponds to y - 21. The vectors defined in 
the above section are given in figure 2a, one row for each vector (for 
each value k). As is immediately seen, the maximum component of the 
- 643 - 
 
	        
Waiting...

Note to user

Dear user,

In response to current developments in the web technology used by the Goobi viewer, the software no longer supports your browser.

Please use one of the following browsers to display this page correctly.

Thank you.