An impulse noise removal filter based on Gaussian mixture distribution model for image

摘要

In this paper, an impulse noise removal method for images is proposed. This method is a switching filter using adaptive Wiener filter based on the Gaussian mixture distribution model in the interpolation process. The adaptive restoration filter is designed as the Wiener filter that minimize mean square error for the Gaussian stationary processes adopted to the local image signals. A simulation result shows that proposed method improves S/N ratio from 1.5 to 2.5 dB compared with conventional switching filter using median filter in its restoration process.