Multiplicative noise and blur removal problems have attracted much
attention in recent years. In this paper, we propose an efficient
minimization method to recover the blurred and noisy image. We make use
of the logarithm to transform multiplicative problems into additive
problems and then employ -norm to measure the data-fitting. The
total variation is also used as a regularization to the recovered image.
We use the alternating direction methods(ADM) to handle the optimization
model. As the set of feasible solutions is nonconvex in the formulation,
we propose to use approximation to make it to be convex, and therefore
make sure the convergence of the proposed algorithm. Experimental results
are report to demonstrate that the proposed algorithm performs better
than the other existing methods.