Filtering the Continuous Wavelet Transform Using Hyperbolic Triangulations

We propose a methodology that detects signal components of a signal in white noise based on a hyperbolic triangulation of its wavelet transform (WT). The theoretical background is a connection between analyticity inducing wavelets and Gaussian analytic functions. This relation allows us to obtain some useful details on the random distribution of the zeros of the wavelet transformed signal. We apply our method to some acoustic signals and observe that many signal components are found but as predicted by the theory there is no guarantee to find all signal components.