Improved Algorithm for Second Generation Wavelet Transform and Image Denosing.

An improved algorithm for second generation wavelet transform is proposed by adopting the lifting scheme and least square orthogonal polynomial fitting. For designing predicting (or updating) coefficients of the lifting scheme, the orthogonal polynomials are taken as a basis, and the predicting (or updating) coefficients are solved by using least square curve fitting to fit the apoximate signals (or detail signals) of wavelet transform. The numerical experiment shows the method is a powerful method for denoising image, which combines second wavelet transform based on optimum fitting estimation and image nonlinear enhancement algorithm, but also it can compensate the foible of soft threshold denoising and keep the feature of image edge well. The proposed method is proved to be effectivly. [ABSTRACT FROM AUTHOR]