New algorithms for recovering highly corrupted images with impulse noise

In this work, we present a new method of noise removal which is applied on images corrupted byimpulse noise. this new algorithm has a good trade-off between quantitative and qualitative properties ofthe recovered image and the computation time. in this new method, the corrupted pixels are replaced byusing a median filter or, they are estimated by their neighbors' values. our proposed method shows betterresults especially in very high density noisy images than standard median filter (smf), adaptive medianfilter (amf) and some other well-known filters for removing impulse noise. experimental results showthe superiority of the proposed algorithm in measures of psnr and ssim, specifically when the image iscorrupted with more than 90% impulse noise.