An efficient reconfigurable framework for impulse noise removal in highly corrupted images

Abstract

newline In the field of digital Image Processing noise removal becomes one of the newlinesignificant process. In DSP the noises are occurs during the occasion of image newlineacquisition otherwise at the time transmitting the image, from this the impulse newlinenoise is mainly affects the image. In this the Impulsive noise is categorized into newlinetwo types such as RVIN and SPN. From these two types of noises, for removing newlineSPN is simple, when compared with the RVIN because of its features. In this newlineimpulsive noise determination is one of the significant filtering process, in this newlineefficiency is increased by the process of techniques applied in the estimation. In newlinethis research work several enhancement techniques are applied to enhance the newlineefficiencies of the impulsive noise determination. For estimating the noise the newlineorientation, directional sub-windows or kernels behaves an important role and it is newlineused to detect the directivity of the noise distribution. In this the decision making newlinesystems will find the pixel corrupted with noise on the basis of output delivered newlinethrough the individual kernel. Then the estimation process of noise removal and newlinethe assurance levels are stored for extra processing. By this process the efficiency newlineof the estimation process is improved through the directional kernels. In this for newlineestimating the noise MAD and S-estimate are used. By using this two technique newlinethe kernels are calculated based on the decision making procedures. The newlineexperimental results show better performance. Here the estimation efficiency is 90 newline% higher than the highly corrupted images. newline