DYNAMIC INTELLIGENT MEAN FILTER FOR IMPULSE NOISE SUPPRESSION IN 2D I MAGES
Kwame Nkrumah, Kwame Osei Boateng · 2013
Noise, as random, unwanted data, appears in images from various sources. So its reduction or removal is an important task in image processing. This study employed an intelligent mean filter, which is an extension of the median filter, adap- tive median filter, and the mean filter, to achieve the pur- pose of noise reduction. The proposed mean filter belongs to the broad class of nonlinear filters. The method is more effective in image processing because it utilizes an intelli- gent technique to find the mean of a set of pixels in the ac- tive window, which is used to perform the filtering process. Impulsive noise is a form of image corruption where each pixel value is replaced with an extremely large or small val- ue that is not related to the surrounding pixel values by a significant probability. Any pixel that is noisy is replaced with the computed mean. The adaptiveness of the method lies in the size of the filtering window, which is determined by the amount of noise in the window. The filtering process starts with a 3x3 window and extends it to a 5x5 window until it gets to the maximum window size chosen by this technique, which is 9x9. The window size extends from the initial size to the subsequent sizes if the amount of noise pollution in each chosen size is greater than 40% using some approximation schemes. The moving-window architecture is employed to the movement of the window through the entire image in order to aid the filtering process. The performance of the proposed intelligent Mean filter has been evaluated in MATLAB simulations on an image that was subjected to various degrees of corruption with impulse noise. The re- sults demonstrated the effectiveness of the algorithm.