A ROBUST OUTLIER DETECTION BASED FILTERING FOR NOISE REMOVAL IN GRAYSCALE IMAGES

Abstract

Salt-and-pepper noise is a common form of impulse noise that severely degrades the visual quality of digital images. Traditional filtering techniques, such as median filters, adaptive median filters, trimmed median filters, and partial differential equation (PDE)-based methods, have been widely used to address this problem. However, these methods often lead to excessive smoothing, loss of fine details, or require complex parameter tuning, especially at high noise densities. This paper presents a novel denoising framework that combines rank filtering, interquartile range (IQR)-based statistical outlier detection, and adaptive image sharpening. Unlike existing techniques, the proposed method dynamically identifies and excludes outliers based on local intensity distribution using IQR analysis, rather than relying on fixed trimming or static window-based rules. This dual-layer approach provides greater robustness against high-density noise while preserving structural details. Experimental results demonstrate that the proposed method achieves superior performance in terms of PSNR, MSE, and SSIM compared to conventional filters, confirming its effectiveness for high-density salt-and-pepper noise removal.