Operator Description of Image Wavelet Denoising
Ke Lin · Harbin Ligong Daxue xuebao · 2000
The scheme of image denoising based on two -dimensional discrete wavelet transform is suggested. The denoising algorithm is described with some operators. By thresholding the wavelet transform coefficients of noisy images, the original image can be reconstructed correctly. Different threshold selection and thresholding methods are discussed. A new adaptive local threshold scheme is proposed. Quantifying the performance of image denoising schemes by using the mean square error, the performance of the adaptive local threshold scheme is demonstrated, and compare with the universal threshold scheme. The experiment shows that image denoising using the former scheme performs better than the one using the latter scheme.