A Decision Based Unsymmetrical Trimmed Modified Winsorized Geometric Mean for the Removal of High Density Salt and Pepper Noise in Images
European Journal of Molecular & Clinical Medicine,
2020, Volume 7, Issue 11, Pages 2256-2263
Abstract
A Decision Based Unsymmetrical trimmed modified Winsorized Geometric mean is proposed for heavily corrupted images. The proposed Algorithm uses a Fixed 3x3 window for increasing noise densities. The algorithm replaces the corrupted pixel with mean of the 4 neighbours or Unsymmetrical trimmed Modified Geometric mean or mean of the window. The use of modified Winsorized Geometric mean provided a better noise elimination characteristic than many Existing algorithms. The Algorithm Exhibits very good noise eliminating capability with improved edge preservation even at very high noise densities.
Keywords:
- Article View: 143
- PDF Download: 188