An Efficient Filter To Remove Universal Noise In High Noise Density Images

Abstract

Digital Images are generally corrupted by Impulse noise during the image acquisition process, while Gaussian noise is encountered during transmission. These noises seriously affect the quality of images. It causes degradation of image spatial resolution, loss of image details and distortion of important image features. Therefore it is essential to correct corrupted pixels before using them in any applications. There are numerous approaches have been proposed to reduce these noises independently. Recently, a Switching Bilateral Filter algorithm is proposed which filters both noises using a single filter but with parameters different for Impulse and Gaussian noise.

 Switching Bilateral Filter (SBF) performances poorly for Impulse noise densities beyond 35%. In this thesis, the above algorithm has been modified to detect impulse noise even at high noise densities. The proposed filter is found to yield better quality images in terms of subjective quality and PSNR values compared to Switching Bilateral Filter.