Explicit iteration schemes for minimization problems arising from image denoising (Q643648)

The author considers explicit iteration schemes for minimization problems arising from image denoising. Some basic facts from convex analysis are reviewed. The author employs the split Bregman method and proposes an explicit iteration scheme based on matrix splitting. The convergence of the proposed iteration scheme is analyzed. When the matrix splitting is done by the symmetric Gauss-Seidel method then the convergence of the scheme is established with no restriction on the step size of the iteration. If the matrix splitting is done by the Gauss-Seidel method then the iteration scheme is shown to converge provided the step size of each iteration is sufficiently small.