Removal Of Low-DENSITY Impulse Noise From Grayscale Images Using New Modified Hybrid MedianFilTerWith Various Sliding Window Sizes

Information is mostly in the form of image and video. One of the major problems in image acquisition, transmission and storage is impulse noise. It is classified into fixed-valued (FVIN) and random-valued (RVIN). Hybrid median filter (HMF) and Modified-HMF are two existing methods in spatial domain to eliminate impulse noise. These methods are modified versions of classical method median filter and applied on low-density RVIN circumstance. In this work, a new spatial domain algorithm New-MHMF is introduced with various sliding window sizes (3×3, 5×5 and 7×7). It is examined both RVIN and FVIN situations. The achievement of NMHMF compared with MHMF by calculating attainment measures Mean Square Error (MSE) and Peak Signal to Noise Ratio (PSNR).