Image restoration based on particle filters

In this paper, a new method for image restoration is proposed based on particle filters, which provides a simple mid flexible algorithm to compute complex posterior distributions for Bayesian image restoration. We incorporate Genetics Algorithm into particle filtering to introduce diversity on the "population" of particles so that improve the robustness, accuracy, and flexibility of the particle filtering. Several experiments are shown to demonstrate the effectiveness of our method in restoring synthetic and real images.