Orthogonal wavelet transform

Section: Wavelets

Comes from “multiresolution” analysis on a dyadic scale
In practice, an Fast (orthogonal) Wavelet Transform can be computed thanks to filter bank and subsampling
Same filter bank can be used to reconstruct the signal from the wavelet coefficients
A wavelet is then fully determined by two filters (which must fullfiled certain conditions): a low pass filter $g$ and a high pass filter $h$