Investigation on salt and pepper noise removal using filtering technique and optimization algorithm

Abstract

Digital images play a crucial role in various fields such as medicine, newlinephotography, biology, astronomy, industry, and defense. Consequently, they newlinedraw the attention of numerous researchers, particularly those focused on newlinepreserving image features from factors that can degrade quality. Current newlineresearch faces challenges with increasing error rates, dimensions, and newlinechanges in image formats. To address these issues, this study proposes a newlinehybrid optimization-based filtering method. The primary objective is to newlineenhance image quality accuracy through effective algorithms. The research newlinebegins by outlining the entire process, starting with the removal of salt and newlinepepper noise using optimization-based filtering techniques. newlineThe hybrid balancing composite motion optimization, Adaptive newlineSwitching Modified Decision-based Unsymmetrical Trimmed Median Filter, newlineand Forensics-Based Investigation approach (R-SPN-ASMD-UTMF-Hyb newlineBCO-FBIA) are used in this work to offer a complete approach for salt and newlinepepper denoising. At first, input photos are taken from many datasets, such newlineas photographs of automobiles, boat kinds and identification, images of cats newlineand breeds, images of butterflies (40 species), and images of 200 species of newlinebirds. These pictures are pre-processed using the ASMD-UTMF filter. The newlinenoisy detection process can be done with the identification of Local Intensity newlineFluctuation (LIF) of salt and pepper noise (SPN). It shows their structure as newlineextensively darker, extensively brighter, single or isolated dots contrast with newlineother nearby pixels and hence the corrupted pixels are considering as newlineintensity function of local extrema. Therefore, proposed Hyb-BCO-FBIA is newlineutilized to optimize the weight parameters of ASMD-UTMF. The population newlinedistribution and forensic-based investigation is initialized uniformly on newlinesolution space. newline