Image Denoising Using Orthogonal Spline

AbstractWe apply new multiresolution theory based on orthogonal spline to image denoising. Just like spline interpolation, the new theory can well approximate any function. It also integrates symmetry, therefore there is no loss or getting and no phase derivation when one reconstruct signal after decomposition. The new method is applied to image denoising via soft-thresholding. Comparing with commonly used wavelets, the new method can well separate noise and image and shows potential application in improving vision quality and preserving edge information for denoised image