
286 Variational Methods in Image Processing
following semi-implicit finite differ e nc es scheme:
u
p
n+1
i,j
− u
p
n
i,j
△t
= −
T
x
p
(x + u(x ))
i,j
(T 1
i,j
− R
i,j
) − γ
∂D
LM
(u)
∂u
p
+ α
"
(λ + µ)A
p
+ µ
u
p
n
i+1,j
− 2u
p
n+1
i,j
+ u
p
n
i−1,j
h
2
+
u
p
n
i,j+1
− 2u
p
n+1
i,j
+ u
p
n
i,j−1
h
2
!
+(λ + µ)
u
q
n
i+1,j+1
− u
q
n
i+1,j−1
− u
q
n
i−1,j+1
+ u
q
n
i−1,j−1
4h
2
where p, q = 1, 2, p 6= q, and
A
1
=
u
1
n
i+1,j
− 2u
1
n+1
i,j
+ u
1
n
i−1,j
h
2
, A
2
=
u
2
n
i,j+1
− 2u
2
n+1
i,j
+ u
2
n
i,j−1
h
2
.
12.2 Variational image registration algorithm using non-
linear elasticity regularization
The focus of this chapter is on a variationa l image registration model that
uses a nonlinea r elasticity regularization. We have chosen to use the standard