Gradient Descent With Smoothness constraints

Asked Sep 30 '15 at 06:24

Active May 02 '18 at 15:42

Viewed 94 times

I have a noisy image Y and known kernel H. I need to estimate a denoised image X such that it gradient of X is also minimised.

J= ||Y-HX||^2+ Alpha* Smoothness constraint(X);

Smoothness constraint= L1norm(|| Grad(X) ||)

how do i estimate the gradient of second term involving smoothness.

Help me on understanding the update rule for x; I only know the gradient for first term.

edited May 02 '18 at 15:42

Cœur

asked Sep 30 '15 at 06:24

user2527599

You need to tell us the form of the smoothness constraint. – dmuir Sep 30 '15 at 08:55
Smoothness constraint= L1norm(|| Grad(X) ||) – user2527599 Sep 30 '15 at 09:52

0 Answers0