Impulse Noise Suppression through Variable Window Size in Color Images
Ashpreet, Mantosh Biswas · 2021 2nd International Conference for Emerging Technology (INCET) · 2021
In general, impulse noise is induced by bit errors in the course of image processing and transmission. Over the last decades, a growing range of updated median filters have been introduced but have their own shortcomings when detecting and restoring noise. In this article, we suggest weighted median filter approaches for salt and pepper (SAP) noise polluted color images using 3×3 and 9×9 window sizes. In our approaches, the pixel is either "noise free pixel" or "noise pixel" utilizing the detection stage. Directional variations and the amount of "right" pixels in the current filter window will adjust the observed noise pixels. The noise free pixel is retained by applying median filtering on it. Simulation results on the taken images show the proposed approaches as enhanced on the prevailing DAMF and BPDF de-noise methods with regard to PSNR, SSIM, NMSE and Computation Time with pictographic comparative study.