The Research On The Performance Of Manifold Wavelet Basis In Image Denoising

From the eighties of the 20 century, The technology of wavelet analysis has developed rapidly, it has special adventage to deal with no calm signal, so it become an important research direction in the domain of image denoising. In this paper , based on using the threshold shrink procedure to denoise, we discussed how the property of the wavelet basis affect the process of image denoising. The main works are as follows:1 When choosing the Daubechies orthogonal wavelets or Biorthogonal wavelets, we compared the effect of using the soft-threshold or hard-threshold, and we brought forward a near standard to choose the threshold function.2 We discussed how the regularity and the number of vanish moment affect the process of image denoising. And we got that it is well to choose the regularity and the number of vanish moment of wavelet basis moderately according to the smooth degree of image.3 In case of different amount of noise and the feature of image, we compared the diffrence between orthogonal wavelet and biorthogonal wavelet. And propose how to choose the orthogonal wavelet and biorthogonal wavelet according to the amount of noise and the feature of image.