A novel wavelet thresholding method for adaptive image denoising

In this paper we present a novel wavelet-based shrinkage technique in conjunction with the nongaussianity measure for image denoising. It provides an adaptive way of setting optimal threshold for wavelet shrinkage schemes, which have in the last decade been shown to yield promising and superior performance than classical methods such as Wiener filtering. Selection of a precise threshold has always remained a difficult issue and is largely done empirically and many methods consider using a universal threshold, which is known to produce over smoothed images. The proposed method selects the threshold adaptively based on image data and leads to improved results. The method makes use of the nongaussianity of the processed image as the performance measure for selection of a particular threshold. Experimental results are provided, together with comparisons with both Wiener filtering and existing wavelet shrinkage schemes.