INTERNATIONAL JOURNAL OF ELECTRONICS AND COMMUNICATION ENGINEERING & TECHNOLOGY (IJECET)

Noise removal is an important aspect in image processing. However, reduction of noise from images that contain textured regions or which have very fine details is difficult. Gaussian noise is very often encountered in acquired images. This paper presents a method to denoise the images corrupted by Additive White Gaussian Noise (AWGN). The noise removal technique proposed here is based on the modification of the well known bilateral filter, which encompasses both the spatial distance and amplitudinal distance between a centre point and its neighboring points. The proposed filter adapts its strength depending on the amount of local noise present in the image. The proposed technique attempts to smooth the image without deteriorating other visual aspects of the image. Experimental results show that the proposed method performs well with different types of images and for a large range of noise deviation. It is effective in removing additive Gaussian noise present in an image along with enhancing the perceptual quality, by virtue of retaining true edges and performing selective smoothing, of the image both qualitatively and quantitatively.