An improved adaptive median-based filter

based on extremum median filter

is proposed to restore images corrupted by salt and pepper noise

especially for those with unknown noise ratio.There are mainly two improved points in this algorithm:(1) a fuzzy measure is proposed to judge the noise ratio;(2)a gradient operator is introduced to save the thin lines

such as edge information.It gets some useful information by training and simulating.Compared with traditional median filter

ordering threshold switching median filter and improved median filter proposed by Sorin Zoican

our approach is more adaptive and receives better results.