Regularization convex optimization method with l-curve estimation in image restoration

As a solution of avoiding ill-posed problem stem from sparse and large scale blurring matrix which has many singular values of different orders of magnitude close to the origin, in image restoration, Tikhonov regularization with l-curve parameter estimation as convex optimization problem has been proposed in this paper. Also, since the restored image is so sensitive to initial guess (start point) of optimization algorithm, a new schema for feasible set and feasible start point has been proposed. Some numerical results show the efficiency of proposed algorithm in comparison with older ones such as reduced newton method.