Previous: The modified Geman-Yang... Next : The reconstruction algorithm |
|
Minimization of the -log likelihood :
![]() |
is fixed and we only compute
,
the gradient of the criterion w.r.t.
.
is fixed
because for each
we can
find a value of
for which
is optimal.
Estimation and
restoration processes are simultaneous :
(the modified Geman-Yang
sampler is initialized with )
|
![]() - |
![]() ![]()
|
![]() ![]() ![]() Generate 2 Markov chains with the Modified Geman-Yang algorithm : |
![]() ![]() ![]() ![]() ![]() |
![]() We stop the algorithm if |