Noise can be introduced into a digitized image in many ways, starting with the lens of the imaging hardware and ending at the digitization of the captured image. The reduction of noise without degradation of the underlying image has attracted much attention in the past. However, whilst many ``structure preserving'' filters have achieved some degree of success at preserving one dimensional image structure, very few have successfully preserved two dimensional image structure. In this paper, we propose a simple image smoothing scheme for improving quality of noisy pictures. The Proposed method forms a weighted mean of the local pixels, with the weights depending on the difference between the central pixel's value and the value of the local pixels. This scheme employs a 3 by 3 mask. The smoothing operation will clean out noises inside a region without blurring its boundary. Simulation studies show that this method reduces Gaussian as well as impulse noise whilst preserving image structure. The software is developed in Mat lab 7.0. Results of applications to several real world images are presented.
