Wavelet-domain denoising based on a new kind of thresholding function with neighbor dependency

A novel thresholding function with neighbor dependency was proposed, which was similar to hard thresholding function presented by Donoho. The proposed thresholding function that was simple and continuous had a high order derivative. The new function overcame the discontinuous shortcoming of the hard thresholding function and the disadvantage of soft thresholding function, which was the invariable dispersion between the estimated wavelet coefficients and the wavelet coefficients contaminated by noise. At the same time, the clustering characteristics of wavelet coefficients were utilized effectively in new function. That was, the neighboring wavelet coefficients were incorporated into the estimation of wavelet coefficients. Simulation results showed that the proposed denoising algorithm had better visual effect and PSNR performance than many exiting thresholding methods.