Smoothing non-equidistantly sampled data using wavelets and cross validation
Maarten Jansen, Adhemar Bultheel · Lirias · 1998
Coefficient thresholding is a popular method in wavelet based noise reduction. A wavelet decomposition is typically a sparse representation ofnoise-free signals: the essential information is captured by a limitednumber of large, important coefficients, while the main part of coefficients isclose to zero. Replacing these small coefficients by zero is a straightforwardway to reduce noise variance without affecting the noise-free signal too much.Recently, algorithms have been developed for wavelet decompositions ofnon-equidistant samples, using so called second generation wavelets.This paper investigates how to apply these algorithms to reducenoise in signals on a non-equidistant grid.