AN IMPULSE NOISE FILTER

Abstract

Present day applications require various kinds of images and pictures as sources of information for interpretation and analysis. Whenever an image is converted from one form to another such as, digitizing, scanning, transmitting, storing, etc., some of the degradation occurs at the output. Hence, the output image has to undergo a process called image enhancement which consists of a collection of techniques that seek to improve the visual appearance of an image. Image enhancement is basically improving the interpretability or perception of information in images for human viewers and providing 'better' input for other automated image processing techniques. One such degradation of image is the addition of impulse noise in an image. One of the widely known forms of impulse noise is “salt and pepper” noise. Various filtering techniques have been proposed for removing impulse noise in the past, and it is well-known that linear filters could produce serious image blurring. As a result, nonlinear filters have been widely exploited due to their much improved filtering performance, in terms of impulse noise attenuation and edge/details preservation. This dissertation presents a novel and efficient approach to impulse noise detection and filtering. The proposed method applies the adaptive fuzzy rule-based technique to noise detection after selecting the decision boundaries discriminatively and assigning the pixels their appropriate class. Thus the decision map formed is given to filtering process. This new technique can remove the impulse noise (represented using four noise models) from corrupted images efficiently and requires no previous training. Quantitative and qualitative analysis, performed on standard color and gray scale images, shows improved performance of proposed technique over existing state-of-art algorithms. The image is corrupted up to 80% noise density under each noise model and PSNR, MSE, correlation coefficient and delta E color difference are used to compare the results of proposed approach with existing algorithms.