Architecture for wavelet packet transform with best tree searching

Wavelet Packet Transform (WPT) provides good spectral and temporal resolutions in arbitrary regions of the time-frequency plane. Given an additive cost function, a best-tree searching algorithm allows the selection of the best basis for a given signal according to this function. This adaptive choice of the time-frequency tiling benefits most of the applications where the standard Wavelet Transform (WT) has already shown to be useful. Though many specific architectures have been proposed in the literature for the WT, it is not the case for WPT. In this work we present a specific architecture for WPT which implements the best-tree searching algorithm.