Improved threshold function of wavalet domain signal de-noising

In this paper, we construct a threshold function on the existing wavelet threshold functions. Compared with the traditional hard threshold, soft threshold, semi-soft threshold and some improved threshold functions, the proposed threshold function has the advantages of easy to calculate and mathematical properties. At the same time, the proposed threshold function also can overcome the drawbacks of hard threshold with discontinuous function and soft threshold function and other threshold functions with constant deviation in de-noising processing. By the Blocks, Bumps, HeaviSines and Doppler signal simulation experiment, the experimental results show that the proposed threshold function can remove the noise and suppress pseudo-Gibbs phenomena effectively. In visual effect, signal-to-noise ratio and mean square error (MSE) measure, the proposed threshold function is superior to the traditional threshold functions and has certain practical value.