Color image interpolation for impulsive noise removal using interactive evolutionary computing

A new type of digital filter for removing impulsive noise from color images is proposed using interactive evolutionary computing. This filter is realized as a rule-based system with interpolation technique considering the correlation among color components. This filter first detects the pixels on which impulsive noise is added on each RGB channel, based on the rules concerning the peculiarity of the signal value and relationship among color components. Then the noisy pixel on each channel is interpolated by the method proposed by Taguchi et al. Here, the problem is how to set the parameters which represent the rules. Interactive evolutionary computing (IEC) is newly adopted to optimize these parameters, considering the subjective assessment of the image quality by human vision. The parameters to judge the degree of peculiarity and the relationship among color components are effectively optimized by IEC and accordingly noise removal can be realized with high quality. Results of computer simulations for noisy images are shown to verify the high performance of the proposed method.