MODIFIED SPATIALLY ADAPTIVE DENOISING ALGORITHM FOR AN IMAGE CORRUPTED BY GAUSSIAN NOISE

Abstract: A modified spatially adaptive denoising algorithm for a single image corrupted by Gaussian noise is proposed in this paper. The proposed algorithm use local statistics of a selected window i.e. by defining local weighted mean, local weighted activity and local maximum. These local statistics are used to detect the noise in the image then a modified Gaussian filter is used for noise suppression. This algorithm is tested against different images and the experimental result shows its result is better than different existing methods like Pixel Wise Median Absolute Difference (PWMAD), Rank Order Criteria (ROC), Switching-based Adaptive Weighted Mean (SAWM) and Spatially Adaptive Denoising Algorithm (SADA).  

Keywords: Gaussian Noise, Denoising, Local statistics and Gaussian filter.