Nonlocal Variational Methods in Image Restoration 91
Exercise 4.3 Show (as a version of the classical divergence theorem):
ˆ
Ω
div
w
~vdx = 0.
Exercise 4.4 Show that the n onlocal Laplacian △
w
is self adjoint:
h△
w
u, ui = hu, △
w
ui.
Exercise 4.5 Show that the n onlocal Laplacian △
w
is negative s emidefinite:
h△
w
u, ui = −h∇
w
u, ∇
w
ui.
Exercise 4.6 Using the not ations from this chapter, consider the difference-
based regularizing functional
Φ(u) =
ˆ
Ω×Ω
φ
(u(y) − u(x))
2
w(x, y)
dydx.
Assuming that φ is differentiable, compute the Euler–Lagrange equation asso-
ciated with the minimization with respect to u of
Φ(u) + λkf − uk
2
L
2
(Ω)
.
Exercise 4.7 In two dimensions, let f be a given noisy image degraded by
additive Gaussian noise of zero mean. Apply the following minimization ...