Decision Based Trimmed Adaptive Windows Median Filter

In this research we have proposed a median filter which is capable of identifying and eliminating salt and pepper noise within the image and keeps the uncorrupted pixels intact. It is based on growing window concept and makes use of simplified and full window for low and high noise densities respectively. The proposed filter replaces the center processing pixels value with the median of the window calculated among noise free candidates. The algorithm works well for both low and high noise density.