IMAGE-RESTORATION USING BIORTHOGONAL WAVELET TRANSFORMS

One well-known method to solve ill-posed problems in image restoration is to use regularization techniques. A new scheme is proposed for digital image restoration based on a regularization method and on the biorthogonal wavelet transform. We show that, in cases where the blur filter can be considered as a low-pass filter of a biorthogonal multiresolution analysis, it is possible to obtain an efficient family of regularization operators from the convolution, operator alone.