Adaptive fuzzy weighted average filter for synthesized image

Monte Carlo is a powerful tool for the computation of global illumination. However noises, which are resulted from the low convergence of Monte Carlo, are noticeable for the synthesized image with global illumination effects and can be regarded as the combination of Gaussian noise and impulse noise. Filter is a cheap way to eliminate noises. In this paper, we investigate nonlinear filtering techniques to reduce the mixturing noise. Based on fuzzy theory, we present an adaptive weighted average filter to optimize the weighs of the filters. Analysis and the computational results, which have been obtained from experiments for noise attenuation and edge preservation, indicate that the new algorithm is promising.