A new filter for removal of salt and pepper noise

In this paper, we have proposed a new method for removal of salt and pepper noise from gray scale images. In this technique when the processing pixel is uncorrupted then it is left unchanged otherwise the neighbors are checked. When all the neighbors are corrupted i.e. having values 0´s or 255´s or both then we go for the mean of the window and the processing pixel is replaced with this mean value. When some of the neighbors are corrupted then we go for unsymmetric trimmed mean value. Simulations are carried out for different percentage of noise density. Experimental results shows that the proposed algorithm provides better result than the Median Filter (MF) and its variants in terms of Peak signal to Noise Ratio (PSNR) and Image Enhancement Factor (IEF).