A Novel Technique for Delineation of Non Linear Noise and Performance Comparison with FPGA Implementation

Abstract

Image enhancement is an essential step in any image processing application. The newlineprocessing of images in medical applications may involve identification of regions of newlineinterest in X-Ray or Computed Tomography (CT) images where it may contain newlineinformation of disease or any organ. In any image processing method, it is required to newlinehave a good quality image to get better results where atomization of image processing newlineis done. Sometimes images may contain noise which may affect the system newlineperformance in applications where image processing atomization is done. The effect of newlinenoise may hide information content and further steps involved in applications of image newlineprocessing may get false results. newlineThe noise removal is a fundamental step in image enhancement. Linear and newlinenonlinear noise are the two main types of the noise. By considering the class of this newlinetype of noise the respective algorithm and mathematical model can be designed for newlineremoval or delineation of the noise. Linear noise can be modelled mathematically and newlinehence fixed filter types can be designed for linear noise delineation. In case of nonlinear newlinenoise, noise detection and delineation is a challenging task due to irregular distribution newlineof noise pixels in the image. In this research work, the problem of linear and nonlinear newlinenoise removal is addressed and solution is provided with the design of filters with multi newlinestage processing that can mitigate nonlinear noise efficiently. newlineIn the first phase of research, the noise delineation using median filters is performed newlineand results on standard images are evaluated. The blurring effect minimization is newlineconsidered as the main side effect during noise removal filtering and block matched 3 newlinedimensional (BM3D) filtering method is used. The BM3D filtering also results in a newlineblurring effect while removing the noise. The image enhancement with noise removal newlinefiltering and then resolution enhancement is performed in steps. The wavelet newlinedecomposition of the image is done in which each part out of four parts is enhanced newlineusing bi-linear inter