Novel filtering methods to improve noise Reduction in digital color images

Abstract

The key interest of this thesis is to develop random valued impulse noise reduction methods for digital color images. The conventionally used vector median and directional the most essential pixels at right angles to the test pixel. This approach overcomesthe problem of color artifacts arising due to joint processing of pixels.As a result of the correlation maintaining characteristic of these filters, the smoothing applied to one colour component smears into the others, creatingcolour artefacts inside the test pixel. This thesis presents a unique approachfor separating the joint vector pixels and performing median filtering on just filters process the pixel vector jointly in the red,green and blue intensities of a centre test pixel in a chosen window of pixels. newline newline