Immune memory clonal selection algorithms for designing stack filters

Stack filters are a class of non-linear filters for suppressing the noise that is uncorrelated with the signal. Their design is formulated as a highly nonlinear optimization problem. A modified immune clonal selection algorithm, called immune memory clonal selection algorithm, is employed to perform the configuration of filters design. The new algorithm has the advantage of preventing from prematurity and fast convergence speed. As an experiment, the stack filters are used to restore images corrupted by uncorrelated additive noise with the level from 10% to 50%. The filters are trained on the small regions of the noise-free and noisy image and then applied to the whole image. The new algorithm has faster convergence speed than that of genetic algorithm. The results are compared with that using the median filter. It turns out that, with our proposed algorithm, a smaller MAE for all noise levels is achieved and much detailed information of the images is preserved. The results show that the new algorithm is effective and feasible.