Competence Investigation of Noise Detecting Procedure Found on TTSD (Triple Threshold Statistical Detection) filter for SPN

Abstract

This paper aims to investigate the TTSD (Triple Threshold Statistical Detection) filter because the TTSD filter, which has been proposed since 2018, is one of the highest recent powerful noise eliminating procedures because an auxiliary condition is first implemented in this procedure which satisfactorily solves the disadvantage of previous noise detection procedures. Moreover, the noise signature is estimated for all pixels and statistically checked with the first threshold to classify noisy pixels succeeds by the statistically checking of the computed pixel with the 2nd and 3rd threshold levels where these thresholds are based on both Gaussian statistical dispersion (mean and standard deviation) and quartile dispersion (median). In this simulation part, many contaminated images, which are comprised of Resolution, Girl, Lena, Pepper, Mobile and Pentagon under salt and pepper noise at many noise densities in the noisy detecting accuracy, noiseless detecting accuracy and overall detecting accuracy perspective.