Image pre-processing is a crucial procedure employed to refine images, making them more visually appealing to humans and facilitating more effective information extraction. Several state-of-the-art methods have been proposed for the purpose of removing noise from images, particularly those afflicted by impulsive noise. This research centers on the examination and evaluation of various noise reduction algorithms, specifically those rooted in the median filter and its advanced nonlinear counterparts. The research also emphasizes the methodologies for tackling impulsive noise de-noising with machine learning techniques. Furthermore, the study sheds light on the shortcomings of one particular approach and suggests potential remedies by considering alternative methods. The evaluation of various state-of-the-art algorithms involves assessing their performance using metrics such as root mean square error (RMSE), peak signal-to-noise ratio (PSNR), and structural similarity index measure (SSIM). The Bayesian selective median filtering (BSMF) method shows a percentage increase in PSNR (average of approximately 14.42%), the Switching Filter with Adaptive Rank Weights (ARWSF) method shows a percentage increase in SSIM (average of approximately 17.645%), and the Convolutional Neural Network (CNN) method shows a percentage decrease in RMSE (average of approximately 15.88%) compared to their counterparts.