Improved Papoulis-Gerchberg algorithm for restoring lost samples

The iterative algorithm of Papoulis-Gerchberg is famous for solving the lost samples recovery problem, however, is usually slowly convergent. In this paper, we propose an efficient approach for restoring lost samples with a preprocess for meeting boundary conditions in the iteration method. The simulation indicates the mean square error (MSE) of the recovery and the convergence rate with the preprocess concept is much better and faster than that without preprocess concept. The improved scheme can also be applied to other cases of signal restoration, which admit Cadzow 's iterative processing method