IMPROVING MEDIAN AND AVERAGE DIGITAL FILTERS PERFORMANCE

Denoising gray and color images from the effects of salt and pepper noise is a vital issue, because clean images
are needed to implement various computer applications. In this paper research a modified median and modified
average filters will be proposed. These filters are to be used to enhance the quality of the denoised gray or color
image, even if this image is infected by a salt and pepper noise with high noise ratio. It will be shown that the
proposed filters will be efficiently used to reduce the negative effects of salt and pepper noise with any noise
ratio (low or high).
The proposed filters will be used to treat only the infected noisy pixels in the noisy image, a special window will
be created and used to calculate the new optimal value of the noisy pixels. This window can be easily created
using simple data structures which include padded noisy image and index matrices. The index matrices will be
used to simplify the process of covering window creation.
The proposed method will be tested using noisy images with various noise ratio values, the results will be
compared with average and median filters results to show the improvements provided by the proposed filters.